ArticlePDF Available

Demographics of Adult Heritage Language Speakers in the United States: Differences by Region and Language and Their Implications

Authors:

Abstract and Figures

Heritage language (HL) speakers have received scholarly attention in recent years as an interdisciplinary research theme, but relatively less attention has been paid to their demographics. ExistingIntegrated Public User Microdata Series (Ruggles & Sobek, 1997), which is based on data from the U.S. Census and the American Community Survey, this study makes geographical and chronological comparisons among groups of adult HL speakers from 1980 to 2010. The data show major differences in the demographics of adult HL speakers in different regions. The analyses also reveal differences by language, specifically between the adult HL speakers of Spanish and other languages. Implications of these patterns are discussed.
Content may be subject to copyright.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
1!
Demographics of Adult Heritage Language Speakers in the United States: Regional and
Typological Differences and their Implications
Tomonori Nagano
LaGuardia Community College
The City University of New York
Education and Language Acquisition Department
31-10 Thomson Avenue (B-234FF)
Long Island City, NY 11101
Email: tnagano@lagcc.cuny.edu
ABSTRACT
Heritage language (HL) speakers have received scholarly attention in recent years as an
interdisciplinary research theme, but relatively less attention has been paid to their
demographics. Existing studies of HL speakers' demographics often focus on young children in
areas of high immigrant concentration (i.e., California, Florida, and New York), and no study has
systematically investigated cross-regional and chronological demographic patterns of adult HL
speakers. From the perspective of HLs as a national resource, such demographic data on adult
HL speakers are useful to gauge the availability of the bilingual workforce and determine
structures needed to support and develop a bilingual U.S. population. Using the Integrated
Public User Microdata Series (Ruggles & Sobek, 1997), which is based on data from the U.S.
Census and the American Community Survey, this study makes geographical and chronological
comparisons among groups of adult HL speakers from 1980 to 2010. The data show major
differences in the demographics of adult HL speakers in different regions. The analyses also
reveal typological differences, specifically between the adult HL speakers of Spanish and other
languages. Implications of these patterns are discussed.
Keywords: heritage language; demographics
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
2!
Heritage language (HL) speakers have drawn much scholarly attention in recent years as an
interdisciplinary research theme among language educators, linguists, and policy makers. For
language educators, understanding how to teach mixed classes with HL speakers and non-
heritage second language learners in their foreign language classrooms is a perennial issue. It is a
challenge because these two groups of language learners have different instructional needs and
goals for learning the target language. Unless the curriculum is specifically developed for HL
speakers, their unique linguistic and instructional needs are often overlooked in the language
classroom. According to linguists, HL speakers contribute to establishing theoretical links
between studies on first language acquisition (FLA) and second language acquisition (SLA).
Research findings on HL speakers, who have undergone the acquisition processes of both first
language learners and second language learners, shed light on the similarities and differences
between FLA and SLA. To policymakers, HLs are national resources for the country, and their
maintenance requires proactive educational programs, such as bilingual programs, and
multilingual public services, such as bilingual ballots.
Over the last two decades, a considerable number of studies have been conducted on HLs
from a wide range of research orientations and broad perspectives. However, in contrast with
these active areas of the HL studies, relatively little is known about the demographics of HL
speakers. Knowledge about the HL speakers’ demographics is often based on anecdotal and
experiential facts from areas with a high immigrant concentration, which tend to have a larger
number of HL speakers (such as California and Florida). The dearth of studies making cross-
regional comparisons and chronological comparisons of HL speakers is particularly prominent.
A purpose of this article is to provide systematic and detailed demographic descriptions
of adult HL speakers in the United States using data available from the U.S. Census and the
American Community Survey (ACS) from 1980 to 2010.
IMPORTANCE OF THE STUDY
A careful study about the demographics of adult HL speakers is necessary to demonstrate
the significance of HL research and to understand the magnitude of benefits that the research
findings bring to the general public. Such information is also crucial for policymakers and
educational institutions to make decisions about bilingual services, such as creating bilingual
ballots, establishing school curricula, and developing language programs and community
services that reflect regional linguistic diversity. Finally, the data will be helpful to gauge the
current availability of bilingual workforce, whose importance has increasingly been emphasized
as the world economy has become highly integrated globally.
In this study, demographic information of adult HL speakers of nearly forty languages for
each state based on census data from 1980 to 2010 are presented. This study focuses on adult HL
speakers (i.e., HL speakers who are at age of 18 or older) because, from the perspective of HLs
as a national resource (Brecht & Ingold, 2002), adult HL speakers’ demographics directly reflect
the current availability of the bilingual workforce in the United States. In addition, data on adult
HL speakers are also useful to capture the chronological transitions of HL speakers because their
language profiles are much less likely to change than those of young children, which are known
to change to English in a short time period.
In terms of the absolute number of HL speakers, the data support the anecdotal and
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
3!
experiential estimations that a large number of HL speakers reside in areas that receive
immigrants, such as California, New York, New Jersey, Texas, Illinois, and Florida. However,
analyses of the fine-grained distribution of HL speakers at the state/county level reveals marked
regional variations and different degrees of dispersion by both language and region. Over the last
thirty years, the growth rate of HL speakers has been much faster than that of the general U.S.
population. However, the growth rate also exhibits regional variations, and some states have
even had a decline in the number of HL speakers.
The following findings regarding the synchronic demographic patterns and diachronic
changes of HL speakers over the past few decades are discussed in this study:
(1) The number of HL speakers grew at a considerably faster rate (26.98% per decade from
1980–2010) than the average growth of the U.S. population (10.88% per decade from
1980–2010). However, the growth rates radically differ from state to state. In some states,
there was even a decline in the number of adult HL speakers.
(2) Spanish and Chinese remain the two most common groups of HLs in the United States,
and their prevalence has grown rapidly over the last 30 years. In addition, although their
numbers are still small in relation to the absolute number of speakers, new HLs such as
Arabic, Hindi, Dravidian, Vietnamese, Russian, Amharic/Ethiopian, and Tibetan, also
experienced substantial growth. At the same time, languages such as French, German,
Italian, Greek, Yiddish, and Dutch are experiencing rapid declines.
(3) Unsurprisingly, states that are typically considered immigration hubs (e.g., California,
Texas, New York, Florida, and New Mexico) have large numbers of HL speakers.
However, in terms of the proportion of HL speakers in the state’s population, some other
states (e.g., Arizona, Hawaii, Massachusetts, Nevada, and New Jersey) have proportions
of HL speakers as large as those in states with high numbers of immigrants.
(4) The majority of HL speakers are speakers of Spanish, but on the state level, there are
quite a few exceptions, such as Alaska (Aleut Eskimo is the most common HL), Hawaii
(Filipino/Tagalog and Japanese), Louisiana (French), Maine (French), New Hampshire
(French), North Dakota (German), South Dakota (Siouan languages), and Vermont
(French).
(5) Over the last 30 years, the concentration of HL speakers has moved from metropolitan
centers to suburbs or non-metropolitan areas. However, some of the HLs continue to have
specific regional concentrations.
PREVIOUS STUDIES
One of the largest surveys of HL speakers to date is the National Heritage Language
Survey, conducted by the National Heritage Language Resource Center (NHLRC) in 2007–2009
(Carreira et al., 2009; Carreira & Kagan, 2011). The NHLRC survey asked a wide range of
questions, including HL speakers’ language usage and proficiency, transition in the use of their
primary languages between childhood and adulthood, motivation, language contact in the HL,
self-assessment of their HL knowledge, self-assessment of their English knowledge, and two
open-ended questions about their attitude toward the HL. The survey was administered in HL
classes in postsecondary institutions, and 1,732 HL speakers responded to the survey.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
4!
Among other findings, the results of the survey highlight the critical role of country of
birth and age of arrival to the United States in the maintenance of HLs. The results show that
speakers with an earlier age of arrival and those born in the United States have a higher tendency
to show lower self-reported proficiency in their HL. In contrast, those who arrived in the United
States at the ages of 11–13 years old and 14–18 years old reported significantly higher self-
reported proficiency in their HL. Age of arrival also correlates with use of the HL. Whereas
those who arrived in the United States before the age of two years primarily use English to speak
with friends (approximately 70–80% of the time), those who arrived between the ages of 2 and
10 years use English a little bit less frequently (60–70%); instead, they frequently employ a mix
of English and the HL (approximately 30% of the time). Among those who arrived in the United
States after the age of eleven, their use of the HL increases considerably and they use a mix of
English and the HL more than 50% of the time. In addition, the survey revealed several intra-
language differences among different HLs. For example, whereas Spanish and Russian HL
speakers tend to maintain exposure to their HLs even after their arrival in the United States, HL
speakers of Korean, Chinese (both Mandarin and Cantonese), Tagalog, and Vietnamese have
significantly reduced opportunities to use their HLs after their arrival in the United States. For
some of the languages in the latter group (e.g., Korean and Chinese), community heritage
schools and religious institutions play a significant role in maintaining HL proficiency because
their ethnic communities are much smaller and more scattered than the communities of the
former groups. Finally, although all HL speakers rated their HL oral skills higher than their
literacy skills, the two language groups that maintain regular contact with their HL (i.e., Spanish
and Russian) reported a higher self-rated proficiency in their HLs than did the Korean,
Vietnamese, Persian, Tagalog, and Chinese groups.
The NHLRC HL Survey provides valuable statistical data that corroborate intuitive and
anecdotal knowledge about the HL speakers, but it also faces some limitations. One major
problem is that, despite the high number of respondents to the survey, the majority (72%) of
respondents were from one state (California) and, as a consequence, the data may not necessarily
reflect the national trends of HL speakers. For example, the language groups of HL speakers
appear to reflect the immigrant communities on the West Coast. The survey respondents
included Armenian (3.4% of the respondents), Persian (3.5%), Tagalog (6.5%), and Vietnamese
(6.6%) speakers, who may not be found as frequently in other regions. The analyses of the
Census/ACS data reveal many regional differences among HL speakers and suggest that these
language communities are typical HL groups in the West Coast states but not in other regions.
Rumbaut et al. (2005, 2009) reported on another set of major surveys on the HL speakers.
Both Immigration and Intergenerational Mobility in Metropolitan Los Angeles (IIMMLA) and
Children of Immigrants Longitudinal Study (CILS) are large-scale surveys of HL speakers in
Southern California and South Florida. The IIMMLA is a cross-sectional study that sampled
5,000 first through fourth generations of immigrants in their 20s and 30s in the Los Angeles area.
The survey found that the majority of the first-, 1.5-, and second-generation immigrants1 used
their HL at home (97.4%, 92.9%, and 83.5%, respectively), but there were huge declines in the
HL proficiency from the first generation (86.9% speak the HL very well) to the 1.5 and second
generations (46.6% and 36.1%, respectively), and then to the third generation (11.9%).
Proficiency in and preference for English exhibited a reverse pattern. Only 17.7% preferred to
speak English among the first generations, whereas the corresponding figures jumped to 60.7%–
73.4% among the 1.5 and second generations and to 97% among the third generation.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
5!
The IIMMLA survey was supplemented by the CILS, another large-scale longitudinal
survey in San Diego and Miami/Fort Lauderdale (Rumbaut, 2009). The CILS survey was
administered to 5,262 speakers of HLs at the age of 14 or 15 in 1992 with two follow-up
interviews in 1995 (when the respondents were 17 or 18) and in 2001–2003 (when the
respondents were in their mid-20s). The results of the CILS survey clearly exhibit strong
linguistic assimilation among children of immigrants: The proportion of 1.5-generation children
who spoke English fluently increased from 42.4% (at 14 or 15 years old) to 54.0% (at 17 or 18
years old) and then to 72.4% (in their mid-20s). Notably, fluency in the HLs also showed
improvement: 50.4% of 1.5-generation children spoke their HLs fluently at 14 or 15 years old,
which increased to 49.3% (at 17 or 18 years old) and 55.8% (in their mid-20s). Rumbaut (2009)
noted a clear typological difference in the HLs in terms of the degree of linguistic assimilation.
The respondents who spoke Spanish had a better chance of maintaining their HL (78.3% of them
understood the HL “very well” in their mid-20s), whereas a much smaller proportion of those
who spoke Asian languages maintained their HLs (40.3%). Using the term “linguistic
graveyard,” Rumbaut (2009) noted an extremely strong pressure for linguistic assimilation in the
United States, which practically extinguishes almost all HLs within the three generations of
immigration.
In addition to these large-scale surveys, several small-scale studies have focused on
specific regions or languages. Lopez (1996) conducted a study of languages spoken in Greater
Los Angeles using U.S. Census data in 1980 and 1990. The examination of census data suggests
a daunting rate of language attrition among 1.5-generation (i.e., those who came to the United
States before the age of ten) and second-generation immigrants, particularly those who speak
Asian languages. According to Lopez (1996), 30% of second-generation Spanish speaking
immigrants speak only English at home. In contrast, the percentages of English monolingual
second-generation immigrants among Asian language speakers are noticeably high (77% for all
Asian languages and 65% for Chinese speakers). The data suggest that concentration in terms of
sheer numbers (i.e., the larger a linguistic community is, the slower the language attrition process
becomes) and isolation (i.e., the more self-contained a linguistic community is, the slower the
language attrition process becomes) are the two factors that facilitate successful maintenance of
HL in Spanish-speaking communities.
Rapid language attrition, particularly among Asian school-age immigrant children, was
also reported by Alba et al. (2002), who conducted multivariate analyses of the 1990 census data.
Their study investigated the influence of various explanatory variables on the Anglicization
process, including parental education, self-employment status, exogamous marriage, the
presence of HL speakers in the household, proximity to the ethnic concentration, and the
percentage of the speakers of the HL in each region. The major finding was that Asian HL
speakers (e.g., Chinese and Filipino) tend to become English monolinguals faster than Hispanic
immigrants (e.g., Mexican and Cuban Spanish speakers). Additionally, two explanatory variables
emerged as the most critical factors for HL speakers to maintain their HLs; the first factor is
parental endogamy/exogamy status, and the second factor is geographic location (in the ethnic
community or not). According to their model, when these two factors are in disfavor of the
maintenance of the HL (i.e., exogamy marriage in a non-ethnic community), the third-generation
children will speak only English (completely lose their HL) 95% of the time. However, the
influence of these two factors differs based on the HL groups. For Chinese HL speakers, the
ethnic community matters much less than for Spanish speakers (Mexicans and Cubans) and,
regardless of the presence or absence of the ethnic community near their dwelling, third-
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
6!
generation children of Chinese HL speakers between exogamous parents almost always lose their
HL. For Spanish speakers, on the contrary, the presence of the ethnic community alleviates the
attrition rate of Spanish among the third-generation HL speakers born between exogamous
parents at a significant rate. (68% in the ethnic community will lose their Spanish by the third
generation, whereas the likelihood increases to 97% in non-ethnic communities.)
Garcia and Fishman (2002) investigated nearly a dozen HLs that are commonly spoken in
New York City. The selection of languages strongly reflects the composition of local immigrant
communities, such as Italian, Greek, Yiddish, Haitian Creole, and Caribbean Creole. In this
volume, Pan (2002) conducted a small-scale survey on Chinese dialects in New York City and
found that the majority of Chinese speakers in New York City did not speak Mandarin as their
primary Chinese dialect. The most common Chinese dialect in his study was Min (36.6%),
including sub-dialects such as Minnan and Minbei, followed by Cantonese (31.7%) and Wu
(20.3%). Mandarin, which is typically considered the most popular Chinese dialect, accounted
for a mere 10.4% of the 200 samples that Pan collected in Manhattan Chinatown and Flushing
Chinatown. However, Mandarin is undoubtedly the most commonly spoken Chinese dialect in
New York City. His data imply that these Mandarin speakers in Chinatown speak Mandarin as a
secondary dialect or even as a second language.
In the same volume, Zentella (2002) shows the diversity of Spanish heritage speakers in
New York City, particularly the Puerto Rican community. Although Spanish is considered to be
a typical HL and HL speakers of Spanish are prevalent across the United States, Zentella argues
that each region has a unique composition of Spanish-speaking communities and questions
monotonic umbrella terms such as ‘Latinos’ and ‘Hispanics.’ According to her, nearly a half of
the Spanish speakers in NYC are Puerto Rican (50.3%), and there are large numbers of
Dominicans (18.7%), Colombians (4.7%), and Ecuadorians (4.4%). Mexicans, who typically
reside in the Southwest and California, account for only 3.3% of the total Spanish-speaking
community in NYC. In NYC, Caribbean dialects of Spanish, from such places as Puerto Rico,
Cuba, and the Dominican Republic, are often sociolinguistically regarded as far less prestigious
than the inland dialects of Spanish, from such places as Chile, Argentina, Columbia, and Mexico.
For example, 60% of Puerto Rican immigrants and 80% of Dominican immigrants responded
negatively to the question, “Should the Spanish of your group be the one taught in NYC
schools?” This lower interest in maintenance of Spanish is one of the reasons for the faster rate
of Anglicization among the Caribbean Spanish speakers. According to data about Puerto Rican
families in East Harlem (Zentella, 1997), a significant number of Puerto Ricans had become
English monolingual or English speakers with weak Spanish skills between 1979 and 1990 (23%
English monolingual in 1979 to 71% in 1990), and 17% of Puerto Ricans reported themselves to
be purely English monolingual in 1990. Although no quantitative data was presented, the rate of
language attrition among speakers of more prestigious dialects of Spanish appears to be much
slower than that of the Puerto Rican community.
Finally, chronological comparisons of the HL speakers’ demographics show new trends
in the commonly spoken HLs in the United States. Using the census data and the survey on
foreign language programs in secondary schools conducted by the Center for Applied
Linguistics, Fee et al. (2014) reported recent transition in the demographics of HL speakers in
the United States. The comparison between the census data in 2009 and 2011 shows a consistent
pattern of Spanish as the dominant HL spoken at home in the United States (Nearly 70% of the
HL speakers use Spanish at home.). However, the composition of the rest of the HLs exhibited
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
7!
rather quick changes within a span of just a few years. For example, the languages of the current
immigrant populations, particularly Hindi, Vietnamese, and Arabic, outnumbered the languages
of the former immigration groups (French, German, and Italian, respectively) in 2011. The
growth of Arabic speakers is particularly notable considering the scarcity of Arabic classes in K–
12 schools. Fee et al. (2014) argue the need for a wider range of bilingual and foreign language
programs in K–12 schools, which are limited to only a selected number of languages, specifically
Spanish, French, German, Japanese, Chinese, and Russian.
Changes in the HL speakers’ demographics have also been reported by Potowski (2010)
and her colleagues. Potowski (2010) presented the demographics, history, and available media
(e.g., newspapers and TV programs in the HL) of the twelve most common HLs and Native
American languages in the United States. Similar to the findings of Fee et al. (2014), Potowski
(2010) also noted gaps between the number of students in foreign language classes at post-
secondary institutions and the number of HL speakers in the United States. However, the
influence of the HLs spoken on the college-level language courses was also observed in the
growth rates of enrollment. The languages that had the greatest growth in college-level language
course enrollment from 2002–2006 were those spoken by recent immigrant families (i.e., Arabic
(127% growth), Chinese (51%), Tagalog (37%), and Korean (37%)). Potowski (2010, pp. 17–18)
proposes that a better understanding of HL populations will benefit the U.S. society through “the
appreciation and promotion of linguistic diversity in the USA.”
To summarize, the earlier studies on HL speakers’ demographics suggest a rapid
Anglicization process that eradicates the functional proficiency of the HL by the third generation
and a great degree of diversity in terms of both regions and languages. Additionally, the majority
of the previous studies focused on young children (Fee et al., 2014; Alba et al., 2002). The
current study will focus on adult HL speakers unlike the earlier studies that investigated the
overall demographics of people who speak non-English languages at home (Shin, 2003, 2010;
Ryan, 2013), young children who speak HLs in the context of their Anglicization process (Alba,
2002), and the unmet needs of bilingual programs at K–12 institutions (Fee et al., 2014).
DEFINITION OF HERITAGE LANGUAGE SPEAKERS
A clear definition of the target population is crucial to obtain their demographic
information. Obviously, the demographic data on HL speakers, such as an estimated number of
HL speakers, depends on how we define HL speakers, since the specificity of the definition
negatively correlates with the estimated size of the target population. If HL speakers are defined
with a set of highly specific conditions, such as place of birth, language input during youth,
languages spoken by parents, level of HL proficiency, awareness of the HL culture, and self-
identity, the estimated number of HL speakers will dwindle to a very small number. Conversely,
the estimate becomes extremely large if HL speakers are broadly defined simply as people who
had at least some exposure to a non-English language in the local community or at home. As
several researchers have argued (Wiley, 2014), no one-size-fit-all definition for HL speakers
exists, and definitions should emerge based on the purpose of and the resources available to a
particular study. In this section, I review several definitions for HL speakers that are commonly
adopted in the existing literature and propose one working definition for the current study.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
8!
Definitions of Heritage Language Speakers
There are essentially three different approaches to defining HL speakers. Fishman (2001)
is one of the oft-cited definitions, which classifies HL speakers into three sub-groups based on
the socio-historical perspectives of the United States. In his definitions, HL speakers can be
classified into three sub-groups: indigenous HL speakers (e.g., speakers of Native American
languages), colonial HL speakers (speakers of, e.g., French, Italian, or German), and immigrant
HL speakers (speakers of, e.g., Spanish, Arabic, Chinese, or Japanese). Fishman’s definition of
HL speakers encompasses a historical account of HL speakers in the United States and is rather
broadly defined. His definition presumes that each individual in the United States who speaks a
language other than English can identify his/her historical roots in a community of speakers of
the language.
Valdés (2001) offers another definition of HL speakers, focusing on linguistic and
genealogical attributes of the HL speakers in school programs. According to Valdés, an HL
speaker is “a language student who is raised in a home where a non-English language is spoken”
and “speaks or at least understands the language (. . .) and is to some degree bilingual in that
language and in English” (Valdes, 2001, p. 38). In this definition, HL speakers should have at
least one parent who speaks a non-English language as a native language and should have
obtained some fluency in the HL when they were raised and educated in an English-speaking
environment. Although Valdés emphasizes that an HL speaker could merely have a minimal
passive understanding in the HL because they have received primarily language input in English,
her definition clearly focuses on the linguistic aspect of HL speakers. An HL speaker speaks both
the HL and English; the latter is their primary language because HL speakersdominant
languages often switches from the HL to English during the pre-puberty years due to English-
medium instruction at K–12 school.
Finally, Hornberger and Wang (2008) attempt to define HL speakers from psychological,
sociopolitical, and socioeconomic perspectives. According to Hornberger and Wang’s definition,
HL speakers are “individuals with familial or ancestral ties to a language other than English who
exert their agency in determining if they are HL speakers of that language” (p.6). This definition
attempts to capture the identity issues of HL speakers and departs from the definitions solely
based on linguistic features, such as HL proficiency. For example, some HL speakers,
particularly those whose HL entails political connotations (e.g., Arabic post 9/11) or for whom
the use of the HL places the speakers into a social minority group, may have different
perceptions from other HL groups. Hornberger and Wang’s definition weighs more on the
psychological aspect of the HL, such as the ideology and self-identity between the mainstream
language (i.e., English) and the HL.
It is helpful to consider these definitions of HL speakers/learners on a spectrum of two
opposing extremes (Polinsky & Kagan, 2007; Carreira & Kagan, 2011). On the one side of the
spectrum is the broad definition, for which the term heritage language learner is often used. On
the one side of the spectrum is the broad definition, which emphasizes the cultural and ethnic
links to their HLs and is akin to the definitions discussed above (Fishman, 2001; Hornberger &
Wang, 2008). On the other side of the spectrum is the narrow definition, which focus on
individuals' linguistic proficiency for the purpose of drawing implications for education and
instruction (Valdes, 2001). Since the linguistic advantages of HL speakers/learners are often lost
within the third generation of immigrants, this narrow definition often considers only the 1.5-
generation, second-generation, and a handful of third-generation immigrant HL speakers.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
9!
HL Speakers and Learners of a Language Other Than English
It is also possible to define HL speakers as one of the subgroups of learners of languages
other than English (LOTEs). HL speakers, particularly those who have not attained a near-native
level proficiency in the HL, often mix with non-HL LOTE students who have never been
exposed to the target language until it is taught in the classroom. According to a national survey
(Furman, Goldberg, & Lusin, 2010), the most commonly taught LOTEs at colleges and
universities in the United States are Spanish (864,986 students), French (216,419), German
(96,349), ASL (91,763), Italian (80,752), Japanese (73,434), Chinese (60,976), Arabic (35,083),
Latin (32,606), Russian (26,883), Greek (20,695), Hebrew (13,807), Portuguese (11,371), and
Korean (8,511). These common LOTEs significantly overlap with what we typically think of as
HLs. It is possible that students who are reported here as LOTE learners are actually HL
speakers taking a college-level modern language course. It is not unusual to find some HL
speakers in an intermediate or advanced modern language course who wish to reconnect with
their ethnic heritage roots or plan to satisfy their foreign language requirement using their
linguistic heritage.
One major difference between LOTE learners and HL speakers is their language
acquisition process. HL speakers receive a considerable amount of input in the HL that their
parent speaks at home and/or in the community. On the contrary, LOTE learners who are not HL
speakers usually start learning a language at K–12 schools or post-secondary institutions, and
they receive a highly structured but extremely limited amount of input in the target language.
Due to these different paths to the acquisition, linguistic profiles of LOTE learners and HL
speakers are different such that HL speakers tend to do well in the spoken modes (i.e., listening
and speaking), whereas LOTE learners exhibit a higher proficiency in the written modes of
language (i.e., reading and writing).
Several studies indicate that HL speakers are equipped with qualitatively different sets of
linguistic knowledge from that of LOTE learners. For example, the advantage of HL speakers
phonological skills is attested in experimental studies (Kagan & Dillon, 2003; Montrul, 2002,
2009; Polinsky & Kagan, 2007; Polinsky, 2011). Au et al. (2002) and Oh et al. (2003) compared
HL speakers and LOTE learners in the same level of language courses and found that HL
speakers performed better than LOTE learners on the accent ratings and the production of stop
consonants in Spanish (i.e., [p, t, k, b, d, g]) (Au et al., 2002) and in Korean (i.e., [ph, th, kh, p, t,
k]) (Oh et al., 2003). However, these two groups of language learners exhibit no significant
difference in their morphosyntax (grammar) tasks.
METHODS
Definition of HL Speakers for the Current Study
As previously discussed, in the research literature, HL speakers have been defined in
terms of their ancestral, linguistic, or psychological identity. HL speakers exhibit a wide range of
characteristics and share some similarities with and differences from LOTE learners. In sum, the
following factors play key roles in defining HL speakers.
(1) Ethnic background: According to Fishman’s definition of HL (Fishman, 2001), HL
speakers can be classified based on the ethnic and ancestral connection to the language.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
10!
(2) Language proficiency: This is likely the most prototypical definition of HL speakers.
Language proficiency and parents’ native language are prominently expressed in Valdés’s
definition of HL speakers (Valdés, 2001).
(3) Identity & sociopolitical circumstances: HL speakers can also be defined by their
psychological and sociopolitical circumstances. This definition focuses on HL speakers' self-
identity rather than their linguistic proficiency (Hornberger & Wang, 2008).
(4) Age of arrival / country of birth / parents’ native language: Demographically speaking,
the majority of HL speakers in the United States have parents who speak a language other than
English. If they are not born in the United States, HL speakers typically arrived in the United
States in their youth, before puberty (Carreira & Kagan, 2011).
In establishing a definition of the HL speakers for the current study, the availability of
data turns out to be an extremely important element to consider. The definition of HL speakers
must be as precise as possible, but at the same time, the degree of specificity is confined by the
available data and the purpose of making such definitions. Unless surveys are administered
specifically for HL speakers as a target population (e.g., Carreira & Kagan, 2011; Mori &
Calder, 2013), demographic studies using pre-existing data are limited in the options to identify
HL speakers. In this study, specificity is limited to the data source, that is, questions concerning
race, ethnicity, immigration age, and home language in the U.S. Census and ACS. In the
previous studies using the U.S. Census data, the following variables are employed to identify HL
speakers (all descriptions here are adopted from the Integrated Public Use Microdata Series
(IPUMS) database (Ruggles & Sobek, 1997)):
(1) Race and Hispanic origin: Race responses have several major selections (e.g., White,
Black, American Indian or Alaska Native) and a selection of “Some other race” where the
respondent can identify their own race. Hispanic origin identifies persons of
Hispanic/Spanish/Latino origin and classifies them according to their country of origin when
possible.
(2) Household: Household serial number is an identifying number unique to each household
record in a given sample.
(3) Birthplace: Birthplace indicates the U.S. state, the outlying U.S. area or territory, or the
foreign country where the person was born
(4) Age: Age reports the person's age in years as of the last birthday.
(5) Language at home: Language at home reports the language that the respondent speaks at
home.
(6) English speaking: English speaking indicates whether [and how well] the respondent is
able to speak English.
(7) Year of immigration: Year of immigration reports the year in which a foreign-born
person entered the United States.
For example, Alba et al. (2002) used nativity information in the census data, such as
household, birthplace, and age, to identify the first generation (foreign-born), the second
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
11!
generation (U.S.-born to at least one foreign-born parent), and the third and later generations
(U.S.-born to U.S.-born parents). Using the same data source, Fee et al. (2014) identified HL
speakers by means of the home language variable. Rumbaut (2005) used the age of immigration
variable in the IIMMLA survey data to classify HL speakers, In his study, the first-generation
immigrants are defined as those who arrived in the United States at age 13 years or later, the 1.5
generation as those who arrived in the United States before age 12 years, the second generation
as those who were born in the United States to two foreign-born parents, and the 2.5 generation
as those who were born in the United States to one foreign-born parent.
For the most part of this study, I use the information about age, home language, English
speaking ability, and year of immigration to identify HL speakers in the data. To be more
specific, I selected individuals from the census data who (a) were age 18 or older, (b) spoke a
language other than English at home, (c) were bilingual (i.e., did not select “Do not speak
English” in the question regarding English proficiency), and (d) immigrated at an age no older
than 18 years old if they were foreign born.
During the analyses, it became obvious that geographic variables have a significant
influence on the demographics of adult HL speakers. To present data fine-tuned for different
geographic areas, I also employed the following geographic variables in this study.
(1) State: States are identified according to the Federal Information Processing Standards
(FIPS) coding scheme.
(2) County / Consistent Public User Microdata Area (ConsPUMA): ConsPUMA identifies
the most detailed geographic areas that can consistently be identified across samples from 1980
onward. There are approximately 2,000 Public Use Microdata Areas (PUMAs), which are
geographic areas representing at least 100,000 people not crossing the state boundaries.
Measure
For the most part, the data are presented in simple and straightforward measures such as
raw numbers (frequency), percentages, and averages. However, the Gini index, an index that I
adopted as a measure for diversity, needs some explanation. The Gini index, also known as the
Gini coefficient, is one of the measures of statistical dispersion similar to standard deviation,
range, and quantiles. It is typically used to represent income disparity, but it can also be
employed to represent any distributional dispersion. The theoretical range of the Gini index is
from 0, which indicates equal distribution across all samples (no inequity), to 1, which indicates
the maximum inequity, meaning all value is centered in one sample point. In this article, I use the
Gini index to represent how the HL speakers are distributed in each state. (ConsPUMA are the
sample unit.)
TABLE 1
Hypothetical Distribution of Heritage Languages in Four Counties
Language 1
Language 2
Language 3
Language 4
Total
County A
40
50
200
2000
2,290
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
12!
County B
40
80
100
1,000
1,220
County C
40
30
20
200
290
County D
40
20
10
100
170
Total
160
180
330
3,300
3,970
Average
40
45.00
82.50
825.00
Standard
Deviation
0
26.46
88.08
880.81
Gini index
0
0.28
0.49
0.49
Table 1 shows a hypothetical state that consists of four counties (County A–County D)
and four heritage languages (Language 1–Language 4). As seen in the Total row, each language
has a different number of speakers. Language 4 has the largest number of speakers, whereas
Language 1 has the smallest number of speakers. The sizes of groups of language speakers are
directly reflected in the average (the total number of speakers / the number of counties). For
example, in an equal distribution, there are 40 speakers of Language 1 in each county. Because
the actual distribution of Language 1 is 40 speakers in each county, the standard deviation and
the Gini index, the two measures of dispersion, are zero (no inequity). Language 2 has an
unequal distribution of speakers, and as a result, its standard deviation and Gini index show a
degree of dispersion (i.e., 26.46 and 0.28). Standard deviation is often used as the measure of
dispersion, but it is difficult to make direct comparisons when the sizes of populations are
different. For example, Language 3 has a significantly higher standard deviation than Language
2 (i.e., 26.46 vs. 88.08) due to its more skewed distribution of speakers. However, the high
standard deviation can also be attributed to the larger quantity of speakers of Language 3 (i.e.,
180 vs. 330). For this reason, the standard deviation by itself cannot be used to compare the
differences in degree of dispersion between the two groups. The Gini index, however, controls
for the size of the population and can be used to compare two groups of different sizes (i.e., 0.28
and 0.49). The comparison between Language 3 and Language 4 demonstrates that as long as the
dispersion remains the same, the Gini index shows the same value regardless of the size of the
population.
DATA AND DISCUSSION
The Data Sources
U.S. Census and ACS data from 1980 to 2010 were collected from the Integrated Public
Use Microdata Series (IPUMS) database (Ruggles & Sobek, 1997). The U.S. Census is a
decennial census that is administered among citizens, non-citizen legal residents, non-citizen
long-term visitors and illegal immigrants dwelling in the United States. The questions cover a
wide range of topics, including gender, age, race, ethnicity, family relationships, home
ownership, education, housing, and jobs. Starting in 2005, the ACS (Gaquin & White–Dunn,
2014), an annual sample-based national survey, has been used to supplement information on
education, housing and jobs and to make statistical estimates about the questions addressed in the
decennial census.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
13!
Data: Languages and Races
The primary variable to identify HL speakers in this study is their home languages. Table
2 shows the list of languages reported in this study. It is worth noting that the linguistic typology
adopted by the U.S. Census Bureau is sometimes criticized for its lack of linguistic accuracy. For
example, all dialects of the Chinese language (e.g., Mandarin, Cantonese, Min, Wu etc.) are
consolidated into one umbrella term, Chinese. Dialectal variations in Arabic and Russian are also
not considered in the home language item of the U.S. Census and ACS.
TABLE 2
Language Classification in the IPUMS (U.S. Census and ACS)
Albanian
Serbo–Croatian, Yugoslavian, or Slavonian
Aleut, Eskimo
Slovak
Amharic, Ethiopian,
etc.
Spanish
Arabic
Sub-Saharan Africa
Armenian
Swedish
Athapascan
Thai, Siamese, or Lao
Burmese, Lisu, or Lolo
Tibetan
Celtic
Turkish
Chinese
Ukrainian, Ruthenian, or Little Russian
Czech
Vietnamese
Danish
Yiddish, Jewish
Dravidian
Zuni
Dutch
African Language
English
American Indian Language
Filipino or Tagalog
Near East Arabic dialect
Finnish
Siouan languages
French
Other Balto-Slavic
German
Other East/Southeast Asian
Greek
Other Malayan
Hamitic
Other or not reported
Other Persian dialects
Although there is no inherent link between race and HL, the information obtained from
the race and ethnicity questions in the census is nonetheless useful to grasp the general trends of
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
14!
HL speakers in the United States. There are two separate questions concerning the respondents’
ethnicity in the U.S. Census. Race asks for respondents’ self-reported ethnicity among the five
categories: ‘American Indian or Alaska Native,’ ‘Asian,’ ‘Black or African American,’ ‘Native
Hawaiian or Other Pacific Islander,’ and ‘White.’ Hispanic Origin asks whether the respondent is
‘Hispanic or Latino Origin’ or ‘Not Hispanic.’ Those who select ‘Hispanic or Latino Origin’
may identify more specific Hispanic origins.
Table 3 shows the summaries of Race and Hispanic Origin questions from the U.S.
Census in 1980, 1990, 2000, and 2010. The overall increase in the population between 1980 and
2010 is 10.91%, and the only race whose population increased below the average was White
(8.5%). The one single ethnic group that had by far the largest increase between 1980 and 2010
was Asian and/or Pacific Islander (63.27%). In terms of Hispanic origin, the total non-Hispanic
population shows a lower growth than average (6.85%), suggesting that the growth of the U.S.
population during 1980 and 2010 was largely driven by the Hispanic population. The ancestry
data among the Hispanic population support the rapid growth of the Hispanic groups, showing
two- to three-fold growths in all sub-groups. In sum, these figures corroborate the recent
anecdotal tendencies of increasing HL courses in Spanish and Asian languages, most notably
Mandarin and other Chinese dialects, in the last decade.
TABLE 3
Race/Ethnicity Data from the U.S. Census Data in 1980, 1990, 2000, and 2010
1980
1990
2000
2010
Average
Increase
Total Population
226,862,400
248,107,628
281,421,906
309,349,689
10.91%
White
194,643,640
208,834,225
231,025,096
248,564,897
8.50%
Black
26,679,320
29,821,972
35,950,767
41,380,302
15.81%
American Indian/Alaska Native
1,539,680
1,999,300
2,635,747
2,767,721
22.23%
Asian and/or Pacific Islander
3,779,800
7,211,476
11,366,992
16,074,005
63.27%
Other race, non-Hispanic
219,960
240,655
443,304
562,764
40.19%
1980
1990
2000
2010
Average
Increase
Total Population
226,862,400
248,107,628
281,421,906
309,349,689
10.91%
Not Hispanic
212,087,320
226,270,777
246,217,426
258,620,119
6.85%
Mexican
8,771,800
13,374,999
20,867,722
32,915,983
55.41%
Puerto Rican
2,035,020
2,632,326
3,400,527
4,682,531
32.08%
Cuban
822,120
1,058,497
1,248,064
1,883,599
32.53%
Other
3,146,140
4,771,029
9,688,167
11,247,457
56.93%
Data from Home Language Questions
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
15!
Using the two questions concerning the home language in the U.S. Census and ACS, it is
possible to make two different estimates of HL speakers in the United States. Figures 1 and 2
show the estimated numbers of HL speakers in the ten states with the highest numbers of HL
speakers in 2010. The estimate in Figure 1 uses only the census question about the use of
languages other than English at home, and according to this estimate, there are 59.5 million
people are classified as HL speakers. Among these 59.5 million possible HL speakers,
approximately 15.2 million speakers, or about one quarter of the total number of HL speakers,
reside in California, 8.1 million speakers (13.6%) in Texas, and 5.47 million speakers (9.2%) in
New York.
According to the definition of HL speakers that I adopt for this study, the estimate of the
home language question alone is undoubtedly too broad an estimate because it may include
young children and first-generation HL speakers (recent immigrants) who speak only the HL
without functional English skills. Because this study is primarily interested in bilingual adult HL
speakers in the United States, it is necessary to consider other factors to separate these non-target
HL speakers from the target population of the current study. By combining the information on
home language, age, immigration year, and self-reported proficiency in English, another estimate
of HL speakers can be obtained (Figure 2). This definition follows a narrower definition of HL
speakers such that the HL speakers are “adults who speak English as their primary language with
some exposure to non-English languages at home.” Based on this estimate, there are 28.7 million
possible HL speakers in the United States. The proportion of HL speakers in each state remains
roughly the same: Slightly less than a quarter of HL speakers, or 6.2 million, reside in California,
4.15 million speakers (14.5%) in Texas, and 2.56 million speakers (8.9%) in New York.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
16!
FIGURE 1: The Estimated Number of HL Speakers for Each State and the Proportion of HL
Speakers to the Total Population of All States
FIGURE 2: The Estimated Number of HL Speakers with Some English Proficiency for Each
State and the Proportion of HL Speakers to the Total Population of All States
15,238,853 spks
(25.61%)
8,101,851 spks
(13.62%)
5,475,744 spks
(9.20%) 4,876,435 spks
(8.20%)
2,633,167 spks
(4.43%) 2,447,053 spks
(4.11%) 1,599,610 spks
(2.69%) 1,339,906 spks
(2.25%)
1,213,141 spks
(2.04%)
1,178,064 spks
(1.98%)
Total number of all states: 59,501,011 speakers
0 5,000,000 10,000,000 15,000,000
California Texas New York Florida Illinois New Jersey Arizona Massachusetts Pennsylvania Georgia
State
Number of Speakers
2010 Census 5% Sample: Respondents who speak languages other than English at home
(The absolute number of speakers and the proportion to the total number of all states in the parenthesis)
6,244,801 spks
(21.75%)
4,150,885 spks
(14.46%)
2,559,744 spks
(8.91%) 2,248,038 spks
(7.83%)
1,245,885 spks
(4.34%) 1,148,519 spks
(4.00%) 818,736 spks
(2.85%) 682,147 spks
(2.38%)
664,236 spks
(2.31%)
611,540 spks
(2.13%)
Total number of all states: 28,713,690 speakers
0 2,000,000 4,000,000 6,000,000
California Texas New York Florida Illinois New Jersey Arizona Massachusetts Pennsylvania Georgia
State
Number of Speakers
2010 Census 5% Sample: Respondents who speak languages other than English at home AND speak at least some English
(The absolute number of speakers and the proportion to the total number of all states in the parenthesis)
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
17!
Table 4 shows the numbers of HL speakers in each state at four time periods a decade
apart. The average increase of HL speakers in the U.S. population from 1980 to 2010 is 26.98%
per decade, which is significantly higher than the growth rate of the total U.S. population per
decade (10.88%, according to the U.S. Census). However, the growth of HL speakers is not
consistent across all U.S. states. The states with large numbers of HL speakers tend to have
higher growth rates. For example, California saw a slightly higher increase rate (32.05%) than
the national average. The same pattern was observed in Texas (36.17%), Florida (47.63%),
Arizona (40.07%), and Georgia (94.87%). Most of these states are typically considered the hubs
of immigrants/ethnic communities, so it is not surprising to see such rapid increases. However,
even among the states with large numbers of HL speakers, some states experienced sluggish
growths, such as New York (10.90%), Illinois (20.43%), New Jersey (22.04%), Massachusetts
(15.91%), and Pennsylvania (11.06%). The patterns of growth are highly inconsistent after these
top ten states. Some states, such as Hawaii (0.31%), Louisiana (6.57%), Maine (9.8%),
Montana (0.92%), New Hampshire (4.56%), North Dakota (23.35%), Rhode Island (2.22%),
South Dakota (4.15%), Vermont (0.75%), and West Virginia (0.45%), saw no increase or a
decline in the number of HL speakers.
TABLE 4
The Numbers of HL Speakers of each U.S. State in 1980, 1990, 2000, and 2010
State
1980
1990
2000
2010
Average Increase
Rate per Decade
California
2,869,920
4,865,328
5,605,047
6,244,801
32.05%
Texas
1,652,160
2,406,404
3,278,754
4,150,885
36.17%
New York
1,893,260
2,021,866
2,482,271
2,559,744
10.90%
Florida
700,400
1,055,856
1,613,390
2,248,038
47.63%
Illinois
718,040
844,290
1,110,040
1,245,885
20.43%
New Jersey
634,320
779,636
1,007,534
1,148,519
22.04%
Arizona
302,580
436,956
675,390
818,736
40.07%
Massachusetts
439,240
494,780
608,088
682,147
15.91%
Pennsylvania
486,920
505,422
562,470
664,236
11.06%
Georgia
89,900
174,692
437,755
611,540
94.87%
Virginia
152,320
286,095
425,421
589,677
58.38%
Washing ton
168,960
251,400
419,396
560,799
49.78%
North Carolina
90,800
156,825
358,547
515,301
81.69%
New Mex ico
300,780
360,675
418,283
469,128
16.01%
Maryland
155,200
253,372
359,333
467,767
45.08%
Colorado
204,180
227,745
359,816
433,319
29.99%
Michigan
355,460
364,382
454,591
432,989
7.50%
Ohio
317,020
341,840
395,323
411,065
9.15%
Connecticut
250,500
257,012
305,365
362,613
13.39%
Nevada
49,240
91,559
217,238
316,210
89.59%
Minnesota
155,240
147,154
221,911
281,471
24.14%
Indian a
132,160
155,184
227,702
273,218
28.05%
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
18!
Louisiana
312,920
306,345
282,450
254,745
6.57%
Oregon
86,640
119,664
211,874
253,828
44.99%
Wisconsin
169,920
159,216
210,802
234,977
12.52%
Tennessee
57,120
84,495
163,105
220,546
58.73%
Missouri
94,820
111,350
163,144
202,600
29.38%
Utah
62,020
81,753
153,186
195,397
48.92%
Oklahoma
82,820
106,675
141,510
178,677
29.24%
Hawaii
163,560
170,786
171,955
164,768
0.31%
South Carolina
47,980
76,335
120,668
159,432
49.77%
Kansas
75,620
88,632
124,194
150,161
26.08%
Alabama
46,240
75,248
105,732
132,426
42.83%
Iowa
62,660
64,827
95,770
113,609
23.27%
Kentucky
39,060
48,258
94,962
109,302
45.14%
Arkansas
26,560
37,620
70,814
96,167
55.23%
Rhode Island
99,900
96,928
99,247
93,218
2.22%
Nebraska
50,420
48,762
75,293
81,144
19.63%
Idaho
31,040
34,360
61,429
77,843
38.73%
Alaska
31,680
44,460
51,858
68,312
29.57%
Mississippi
29,180
41,040
62,279
61,843
30.57%
Maine
83,340
73,524
71,988
60,803
9.80%
Delaware
19,260
8,608
38,887
57,954
115.16%
New Hampshire
66,820
58,973
64,345
57,229
4.56%
Washing ton DC
34,340
45,912
58,990
49,005
15.09%
South Dakota
39,600
33,576
30,933
34,218
4.15%
Montana
26,900
25,270
32,166
26,226
0.92%
North Dakota
57,340
34,025
29,824
23,851
24.35%
Wyoming
19,480
18,972
19,903
23,346
6.53%
West Virginia
25,980
23,580
32,328
22,881
0.45%
Vermont
21,500
17,524
21,444
21,094
0.75%
State not identified
0
121,137
0
0
NA
Total
14,083,320
18,736,328
24,404,745
28,713,690
26.98%
Note. Source: The U.S. Census/ACS from the IPUMS.
Table 4 indicates rather inconsistent patterns of HL speakers’ demographics at the state
level, which is corroborated by other types of statistics. The number of HL speakers is one
indicator of the presence of adult HL speakers in each state, but it has an inherent bias toward
larger states. The true representation of HL speakers can be better represented by the proportion
of HL speakers, controlling for the difference in the populations of each state. Table 5 shows
such a statistic, the proportions of HL speakers to the state populations.
TABLE 5
The Proportions of HL Speakers to the State Population for each U.S. State in 1980, 1990, 2000,
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
19!
and 2010
State
1980
1990
2000
2010
Difference
between
2010 and
1980
State
1980
1990
2000
2010
Difference
between
2010 and
1980
Alabama
1.18%
1.87%
2.38%
2.77%
1.58%
Montana
3.40%
3.17%
3.56%
2.65%
0.75%
Alaska
7.80%
8.07%
8.28%
9.57%
1.76%
Nebraska
3.20%
3.09%
4.40%
4.43%
1.23%
Arizona
11.11%
11.96%
13.16%
12.77%
1.66%
Nevada
6.12%
7.66%
10.86%
11.69%
5.57%
Arkansas
1.16%
1.60%
2.65%
3.29%
2.13%
New Hampshire
7.27%
5.34%
5.21%
4.35%
2.93%
California
12.10%
16.37%
16.54%
16.72%
4.62%
New Jersey
8.60%
10.13%
11.97%
13.05%
4.45%
Colorado
7.05%
6.94%
8.37%
8.58%
1.53%
New Mexico
22.96%
23.93%
23.00%
22.71%
0.25%
Connec ticut
8.05%
7.83%
8.96%
10.14%
2.09%
New York
10.78%
11.28%
13.08%
13.20%
2.42%
Delaware
3.26%
1.30%
4.97%
6.44%
3.18%
North Carolina
1.54%
2.37%
4.46%
5.39%
3.85%
Washing ton
DC
5.35%
7.53%
10.32%
8.11%
2.75%
North Dakota
8.74%
5.36%
4.64%
3.54%
5.20%
Florida
7.17%
8.17%
10.09%
11.93%
4.76%
Ohio
2.94%
3.16%
3.48%
3.56%
0.63%
Georgia
1.65%
2.70%
5.35%
6.30%
4.65%
Oklahoma
2.74%
3.40%
4.10%
4.75%
2.01%
Hawaii
16.87%
15.46%
14.19%
12.08%
4.79%
Oregon
3.28%
4.22%
6.19%
6.61%
3.33%
Idaho
3.29%
3.47%
4.75%
4.95%
1.67%
Pennsylvania
4.10%
4.26%
4.58%
5.23%
1.13%
Illinois
6.27%
7.41%
8.94%
9.70%
3.43%
Rhode I sland
10.54%
9.67%
9.47%
8.85%
1.69%
Indiana
2.41%
2.80%
3.74%
4.21%
1.80%
South Carolin a
1.53%
2.19%
3.01%
3.44%
1.90%
Iowa
2.15%
2.34%
3.28%
3.73%
1.58%
South Dako ta
5.71%
4.84%
4.10%
4.19%
1.52%
Kansas
3.20%
3.59%
4.62%
5.25%
2.05%
Tennessee
1.24%
1.74%
2.87%
3.47%
2.23%
Kentucky
1.07%
1.32%
2.35%
2.51%
1.45%
Texas
11.58%
14.20%
15.73%
16.43%
4.86%
Louisiana
7.43%
7.31%
6.32%
5.61%
1.82%
Utah
4.24%
4.75%
6.87%
7.04%
2.80%
Maine
7.41%
5.99%
5.64%
4.58%
2.83%
Vermont
4.18%
3.13%
3.52%
3.37%
0.81%
Maryland
3.68%
5.30%
6.78%
8.08%
4.40%
Virginia
2.85%
4.63%
6.01%
7.35%
4.50%
Massachusetts
7.67%
8.22%
9.57%
10.40%
2.73%
Washing ton
4.08%
5.19%
7.11%
8.31%
4.23%
Michigan
3.83%
3.92%
4.58%
4.38%
0.55%
West Virg inia
1.33%
1.32%
1.79%
1.23%
0.10%
Minnesota
3.81%
3.37%
4.52%
5.30%
1.49%
Wisconsin
3.60%
3.26%
3.93%
4.13%
0.53%
Mississippi
1.15%
1.60%
2.19%
2.08%
0.93%
Wyomin g
4.12%
4.19%
4.03%
4.14%
0.01%
Missouri
1.93%
2.18%
2.92%
3.38%
1.45%
State not
identified
NA
NA
NA
NA
NA
Average of all
states
6.21%
7.55%
8.67%
9.28%
3.07%
Note. Source: The U.S. Census/ACT from the IPUMS.
Unsurprisingly, states with large immigrant hubs tend to have relatively high proportions
of adult HL speakers. For example, 16.72% of California’s state population is comprised of HL
speakers; this figure is 16.43% for Texas, 13.20% for New York, 11.93% for Florida, and 9.70%
for Illinois. Notably, some states that are not ranked high in raw number of HL speakers have
equally large proportions of HL speakers in their population. For example, New Mexico is
ranked the 14th largest state in terms of the raw number of HL speakers, but as many as 22.71%
of their state population is identified as HL speakers. The same pattern is observed in
Connecticut (10.14%), Nevada (11.69%), Hawaii (12.08%), Rhode Island (8.85%), and the
District of Columbia (8.11%). A proportion of HL speakers to the state population is critical
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
20!
information because it has a significant effect on securing each state’s public services for HL
speakers. The high proportion of adult HL speakers should be interpreted as a need for a larger
allocation of public resources for HL speakers in these states, such as bilingual education
programs and bilingual access to public services.
The total number of HL speakers is a useful statistic, but it obfuscates the composition
and density of adult HL speakers, two important aspects of HL speakers’ demographics, and falls
short of demonstrating the complexity of the demographics of adult HL speakers.
Composition of HL Speakers
The raw number of HL speakers does not show the composition of HLs in each region.
As the race/ethnicity data suggest, Hispanics account for a significant proportion of total HL
speakers, and the 2000–2010 period saw an influx of Hispanic and Asian immigrants. This
pattern should be reflected in the composition of HLs in the United States; that is, Spanish
should be the largest HL community in the United States, followed by Asian languages, such as
Chinese. Using the home language data, we can obtain such linguistically fine-grained estimates
of the number of HL speakers in various languages (see Table 6).
TABLE 6
The Numbers of HL Speakers by Language in 1980, 1990, 2000, and 2010
Language
1980
1990
2000
2010
Percentage
in 2010
Average
Increase rate
per decade
Spanish
6,465,280
9,421,958
13,661,063
17,013,399
59.25%
38.42%
French
1,160,560
1,365,599
1,372,867
1,256,193
4.37%
3.23%
Chinese
375,800
776,851
1,005,682
1,233,957
4.30%
52.96%
Hindi and rel ated
163,760
394,952
746,244
1,185,354
4.13%
96.32%
Filipino , Tagalog
354,800
648,741
729,341
789,915
2.75%
34.53%
German
966,920
966,303
893,169
715,471
2.49%
9.18%
Vietna mese
113,800
330,910
485,837
540,677
1.88%
82.96%
Korean
187,700
429,364
445,707
487,432
1.70%
47.31%
Russian
88,000
154,015
449,141
481,380
1.68%
91.27%
Arabic
146,140
239,654
338,353
439,744
1.53%
45.05%
Dravidian
26,900
65,199
194,481
423,649
1.48%
152.83%
Italian
974,940
796,184
578,891
393,700
1.37%
25.87%
Portuguese
206,180
235,376
281,333
345,253
1.20%
18.80%
Sub Saharan Africa
35,820
80,853
183,040
335,973
1.17%
111.89%
Polish
602,880
520,920
430,122
310,821
1.08%
19.59%
Japanese
245,420
295,689
312,001
254,039
0.88%
2.47%
Serbo Cro atian, Yugoslavian,
Slavonian
92,900
78,208
128,736
182,954
0.64%
30.30%
Greek
232,740
215,638
186,452
159,108
0.55%
11.85%
Persian, Iranian, Farsi
80,760
147,809
139,201
146,202
0.51%
27.41%
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
21!
Navajo Navaho
NA
99,298
126,387
139,291
0.49%
18.75%
Thai, Siamese, Lao
56,000
131,887
106,895
108,458
0.38%
39.34%
Amharic, Ethi opian, etc
8,180
26,107
55,837
108,221
0.38%
142.28%
Hebrew, Israeli
60,100
89,757
114,796
106,610
0.37%
23.37%
Indonesian
11,040
29,582
40,959
97,439
0.34%
114.77%
Armenian
64,620
90,430
95,675
95,593
0.33%
15.22%
Micronesian, Polynesian
34,260
41,264
41,411
87,711
0.31%
44.20%
Yiddish, Jewi sh
149,420
111,973
100,281
82,608
0.29%
17.71%
Dutch
81,720
75,921
88,776
80,717
0.28%
0.25%
Albanian
7,000
10,394
43,827
77,777
0.27%
149.20%
Ukrainian, Ru thenian, Little Russian
48,400
45,199
64,437
77,290
0.27%
18.63%
Rumanian
20,140
44,701
68,141
76,918
0.27%
62.42%
Tibetan
5,880
34,069
43,552
75,687
0.26%
193.67%
Other East Sou theast Asian
8,760
64,685
67,974
67,401
0.23%
214.22%
Turkish
18,460
27,618
46,585
66,476
0.23%
53.66%
Magyar, Hungarian
89,240
72,088
56,916
39,480
0.14%
23.63%
Lithuanian
43,780
37,496
28,938
34,534
0.12%
5.95%
Swedish
60,620
49,218
45,134
33,126
0.12%
17.90%
Czech
90,560
67,121
54,993
32,438
0.11%
28.32%
Norwegian
78,860
57,130
38,242
27,090
0.09%
29.93%
Slovak
63,100
66,629
29,751
21,179
0.07%
26.19%
Other o r not reported
185,960
6,993
163,647
18,810
0.07%
685.14%
Finnish
51,540
41,132
28,868
17,999
0.06%
29.22%
Others
110,740
291,879
291,062
445,616
1.55%
72.13%
American Indian, all
213,640
NA
NA
NA
NA
NA
ALL
13,869,680
18,776,794
24,404,745
28,713,690
100.00%
27.67%
By looking at the average increase rates, it becomes obvious that there are two groups of
HL speakers in terms of the average growth rate per decade between 1980 and 2010. On one
hand, speakers of French (3.23%), German (9.18%), Italian (25.87%), Portuguese (18.80%),
Polish (19.59%), Japanese (2.47%), and Greek (11.85%) are undergoing declines in the
number of HL speakers or much smaller growth than the nation’s average growth rate of
27.67%. The other group is HL speakers of languages spoken by recent immigrants such as
Spanish (38.42%), Chinese (52.96%), Hindi (96.32%), Filipino/Tagalog (34.53%), Vietnamese
(82.96%), Korean (47.31%), Russian (91.27%), Arabic (45.05%), and Dravidian (152.83%).
This growth trajectory from 1980 to 2010 predicts that some of the commonly spoken HLs in the
United States (e.g., French, German, Italian, Portuguese, and Japanese) may diminish in the next
few decades. In fact, according to a recent data by Fee et al. (2014), Chinese has already replaced
French and become the second most commonly spoken HL, and Vietnamese has replaced
German as the sixth most commonly spoken HL.2
Additionally, the ranking of the HLs in Table 6 might not match our intuitive estimation.
For example, Filipino/Tagalog, the fifth largest HL group in the United States, is not very
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
22!
common in New York. Italian, on the contrary, is one of the typical HLs in New York, but it is
barely found in West Coast states, such as California. Navajo, the 30th most common HL in the
United States, is one of the most typical HLs in Arizona and New Mexico, but it is nonexistent
outside of these two states. Obviously, there are many regional and geographic variations in the
distribution of HLs in the United States.
To illustrate these regional differences, the fifteen most common HLs in California,
Texas, New York, Florida, and New Mexico are plotted in Figure 3. In California, Spanish
(63.41%), Chinese (6.51%), Filipino/Tagalog (65.56%), Vietnamese (2.80%), and Korean
(2.52%) account for higher proportions than the national averages. On the contrary, there are
fewer proportions of French, German, Italian, Arabic, Hindi, Russian, Portuguese, and Polish in
California.
The composition of HLs in New York is rather stark. One noticeable difference is that the
proportion of Spanish speakers in New York (48.00%) is much lower than the national average.
Additionally, in New York, French (5.87%), Italian (3.83%), Russian (4.33%), and Polish
(1.91%) have higher proportions than the national averages. These languages are less common in
California, exhibiting complementary patterns of HLs between California and New York.
Another interesting pattern observed in the comparisons of HLs in California and New York is
the proportion of Spanish HL speakers. Spanish is the most frequently spoken HL in both states,
but the proportions are different. In California, Spanish HL speakers account for 63.41% of the
total HL speakers, whereas they account for 48.00% in New York. In New York, the HLs other
than Spanish collectively represent more than half of the HL speakers (nearly 50%) in the state.
Texas contrasts greatly with New York in its distribution of HLs because Spanish HL
speakers account for as much as 84.09% of the total HL speakers in Texas. In Texas, Chinese,
which is the second most commonly spoken HL in California and New York, has only a
marginal representation (1.55%) and is outnumbered by Hindi and Vietnamese.
In fact, the comparison of these states makes it clear that the only shared pattern is that
Spanish is the most commonly spoken HL. Other than this single fact, almost everything else,
such as the proportions of HL speakers and the rankings of commonly spoken HLs, radically
differs from state to state. For example, in Florida, Spanish accounts for 70.45% of the total HL
speakers, followed by French (10.28%) and Portuguese (2.20%). In New Mexico, Spanish
accounts for 76.83%, followed by Navajo (11.18%) and Zuni (2.06%). As discussed above, in
New York, Spanish accounts for much smaller proportion (only 48%) than those of Florida and
New Mexico, followed by Chinese (8.04%), French (5.87%), and Hindi (5.75%).
In addition, if we look at all the U.S. states, we can find quite a few exceptions where
Spanish is not the most common HL, such as Alaska (Aleut Eskimo is the most common HL),
Hawaii (Filipino/Tagalog and Japanese), Louisiana (French), Maine (French), New Hampshire
(French), North Dakota (German), South Dakota (Siouan languages), and Vermont (French).
Thus, in sum, this survey of the composition of HLs reveals no consistent pattern at the state
level.
FIGURE 3
The Fifteen Most Commonly Spoken HLs in California, Texas, New York, Florida, and New
Mexico
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
23!
3,959,847 spks (63.41%)
406,451 spks (6.51%)
346,916 spks (5.56%)
213,968 spks (3.43%)
174,798 spks (2.80%)
157,227 spks (2.52%)
86,501 spks (1.39%)
82,802 spks (1.33%)
77,470 spks (1.24%)
71,977 spks (1.15%)
69,194 spks (1.11%)
65,305 spks (1.05%)
62,137 spks (1.00%)
53,592 spks (0.86%)
37,632 spks (0.60%)
Italian
German
Dravidian
Persian Iranian Farssi
Armenian
Arabic
Japanese
French
Russian
Korean
Vietnamese
Hindi and related
Filipino Tagalog
Chinese
Spanish
0 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000
Number of Speakers
Heritage Languages
2010 Census 5% Sample: Respondents who speak languages other than English at home AND
speak at least some English by language (The proportion to the total number in the parenthesis)
California
3,490,634 spks (84.09%)
108,421 spks (2.61%)
74,327 spks (1.79%)
64,432 spks (1.55%)
52,999 spks (1.28%)
47,195 spks (1.14%)
40,929 spks (0.99%)
37,306 spks (0.90%)
34,236 spks (0.82%)
31,043 spks (0.75%)
26,487 spks (0.64%)
13,338 spks (0.32%)
12,287 spks (0.30%)
12,265 spks (0.30%)
9,949 spks (0.24%)
Japanese
Persian Iranian Farssi
Portuguese
Russian
Korean
Arabic
Sub Saharan Africa
Filipino Tagalog
German
French
Dravidian
Chinese
Vietnamese
Hindi and related
Spanish
0 1,000,000 2,000,000 3,000,000 4,000,000
Number of Speakers
Heritage Languages
2010 Census 5% Sample: Respondents who speak languages other than English at home AND
speak at least some English by language (The proportion to the total number in the parenthesis)
Texas
1,228,676 spks (48.00%)
205,918 spks (8.04%)
150,235 spks (5.87%)
147,129 spks (5.75%)
110,821 spks (4.33%)
98,088 spks (3.83%)
63,358 spks (2.48%)
50,700 spks (1.98%)
48,900 spks (1.91%)
42,566 spks (1.66%)
41,438 spks (1.62%)
39,827 spks (1.56%)
35,940 spks (1.40%)
33,801 spks (1.32%)
29,008 spks (1.13%)
Greek
Hebrew Israeli
Arabic
Filipino Tagalog
Sub Saharan Africa
German
Polish
Korean
Yiddish Jewish
Italian
Russian
Hindi and related
French
Chinese
Spanish
0 500,000 1,000,000 1,500,000
Number of Speakers
Heritage Languages
2010 Census 5% Sample: Respondents who speak languages other than English at home AND
speak at least some English by language (The proportion to the total number in the parenthesis)
New York
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
24!
Density of HL Speakers
Another aspect of the demographics of HL speakers is their density in a specific region.
The raw number of HL speakers does not show how dense the population of HL speakers is in
each community. Information about density is particularly crucial when we determine the
distribution of resources for HL speakers because the availability of resources, such as
community language schools, must be located within an accessible, commutable distance from
where HL speakers reside. For example, there are approximately 2.6 times more HL speakers in
California than in New York, but in terms of geographic area, California (155,959 square miles)
is 3.3 times larger than New York (47,213 square miles). Therefore, the raw numbers of HL
speakers may not directly result in a higher density of HL speakers in California than in New
York.
In this study, the Gini index is employed to investigate the density of HL speakers. The
Gini index is a statistic of dispersion and shows the degree of distribution in a range from zero to
one. A Gini index of zero indicates an equal spread (i.e., HL speakers are equally distributed
across all ConsPUMAs/counties), and a Gini index of one indicates the most skewed spread (i.e.,
HL speakers are centered in one single ConsPUMA/county). Table 7 shows the Gini indexes for
all HLs.
1,583,780 spks (70.45%)
231,177 spks (10.28%)
49,491 spks (2.20%)
42,173 spks (1.88%)
40,511 spks (1.80%)
31,314 spks (1.39%)
27,498 spks (1.22%)
26,355 spks (1.17%)
24,609 spks (1.09%)
17,699 spks (0.79%)
17,316 spks (0.77%)
16,174 spks (0.72%)
12,815 spks (0.57%)
12,642 spks (0.56%)
10,978 spks (0.49%)
Hebrew Israeli
Polish
Dravidian
Russian
Serbo Croatian Yugoslavian Slavonian
Arabic
Chinese
Vietnamese
Italian
Filipino Tagalog
German
Hindi and related
Portuguese
French
Spanish
0 500,000 1,000,000 1,500,000 2,000,000
Number of Speakers
Heritage Languages
2010 Census 5% Sample: Respondents who speak languages other than English at home AND
speak at least some English by language (The proportion to the total number in the parenthesis)
Florida
360,410 spks (76.83%)
52,465 spks (11.18%)
9,680 spks (2.06%)
9,385 spks (2.00%)
7,029 spks (1.50%)
4,244 spks (0.90%)
3,906 spks (0.83%)
2,608 spks (0.56%)
2,531 spks (0.54%)
1,950 spks (0.42%)
1,886 spks (0.40%)
1,388 spks (0.30%)
1,099 spks (0.23%)
867 spks (0.18%)
817 spks (0.17%)
Thai Siamese Lao
Russian
Dravidian
Filipino Tagalog
Vietnamese
Athapascan
Arabic
Chinese
French
German
Native
Keres
Zuni
Navajo Navaho
Spanish
0 100,000 200,000 300,000 400,000
Number of Speakers
Heritage Languages
2010 Census 5% Sample: Respondents who speak languages other than English at home AND
speak at least some English by language (The proportion to the total number in the parenthesis)
New Mexico
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
25!
TABLE 7
GINI Index of HL Speakers for Each HL
Language
1980
2000
2010
Change
1980
2010
Language
1980
2000
2010
Change
1980
2010
Spanish
0.828
0.769
0.764
0.063
Rumanian
0.902
0.843
0.840
0.062
French
0.774
0.728
0.766
0.007
Tibetan
0.918
0.940
0.948
0.031
Chinese
0.853
0.841
0.821
0.032
Hamitic
0.935
0.952
0.950
0.015
Hindi and related
0.778
0.783
0.786
0.008
Other East Southeast
Asian
0.911
0.892
0.913
0.003
Filipino Tagalog
0.875
0.854
0.847
0.029
Turkish
0.839
0.813
0.839
0.000
German
0.615
0.569
0.621
0.006
Native
NA
NA
0.929
NA
Vietnamese
0.797
0.835
0.845
0.048
Other Balto Slavic
0.900
0.843
0.902
0.002
Korean
0.798
0.825
0.854
0.055
Magyar Hungarian
0.810
0.754
0.852
0.042
Russian
0.831
0.833
0.808
0.023
Lithuanian
0.853
0.825
0.897
0.043
Arabic
0.767
0.749
0.773
0.006
Near East Arabic
dialect
0.914
0.959
0.972
0.058
Dravidian
0.811
0.798
0.820
0.009
Swedish
0.768
0.754
0.867
0.099
Italian
0.804
0.752
0.765
0.038
Czech
0.835
0.800
0.877
0.042
Portugu ese
0.904
0.851
0.855
0.048
Other Malayan
0.902
0.903
0.940
0.038
Sub Saharan Africa
0.809
0.815
0.816
0.007
Athapascan
NA
0.986
0.953
NA
Polish
0.815
0.810
0.854
0.039
Burmese Lisu Lolo
0.963
0.937
0.927
0.036
Japanese
0.859
0.795
0.822
0.037
Norwegian
0.868
0.794
0.895
0.027
Serbo Croatian
Yugoslavian Slavonian
0.825
0.767
0.837
0.011
Aleut Eskimo
NA
0.997
0.997
NA
Greek
0.791
0.773
0.803
0.012
Slovak
0.885
0.822
0.881
0.004
Persian Iranian Farsi
0.786
0.878
0.900
0.113
Other or not reported
0.595
0.767
0.893
0.298
Navajo Navaho
NA
0.985
0.990
NA
Other Persian dialects
0.933
0.941
0.950
0.017
Thai Siamese Lao
0.771
0.769
0.839
0.068
Finnish
0.874
0.841
0.916
0.042
Amharic Ethiopian etc
0.926
0.903
0.912
0.013
Hawaiian
0.957
0.958
0.980
0.024
Hebrew Israeli
0.882
0.855
0.895
0.013
Siouan languages
NA
0.940
0.985
NA
Indonesian
0.872
0.834
0.867
0.006
Danish
0.775
0.787
0.902
0.126
Armenian
0.931
0.955
0.965
0.034
Celtic
0.770
0.817
0.917
0.147
Micronesian Polynesian
0.919
0.895
0.925
0.006
Zuni
NA
NA
0.998
NA
Yiddish Jewish
0.920
0.912
0.972
0.052
Keres
NA
0.993
0.998
NA
Dutch
0.736
0.676
0.799
0.064
African NS
0.901
0.909
0.967
0.066
Albanian
0.932
0.899
0.912
0.020
Iroquoian
NA
0.914
0.969
NA
Ukrainian Ruthenian
Little Russian
0.825
0.808
0.881
0.056
American Indian NS
NA
NA
0.960
NA
Average
0.844
0.846
0.886
0.026
The average Gini index was 0.886 in 2010, indicating that HL speakers usually reside in
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
26!
specific areas of each state. This is not surprising because it is a typical pattern of demographics
that populations cluster around large cities. The Gini index also shows the different degrees of
distribution across the HL groups. The race/ethnicity data show that Hispanic communities are
more spread out across counties than the Chinese, Filipino/Tagalog, Korean, Portuguese, and
Japanese communities are. The Gini index for Spanish HL speakers was 0.764 in 2010, the
second to the lowest after German HL speakers (0.621). In comparison, HL speakers of Chinese
(0.821), Filipino/Tagalog (0.847), Vietnamese (0.845), Korean (0.854), Dravidian (0.820),
Portuguese (0.855), and Japanese (0.822) are more densely concentrated than are Spanish and
German communities. The transition of the density of HL communities from 1980 to 2010 shows
that the HL community as a whole has moved away from the high-density immigrant enclaves
and spread towards the suburbs and outside of the central cities. This tendency is most obvious
with the Spanish community, whose Gini index reduced by 0.063 from 0.828 in 1980 to 0.764 in
2010.
SUMMARY OF DISCUSSION
As presented at the beginning of this study, the survey of the U.S. Census and ACS data
from 1980 to 2010 reveals the following demographic patterns of HL speakers in the United
States:
(1) The number of HL speakers grew at a considerably faster rate (26.98% per decade from
1980–2010) than the average growth of the U.S. population (10.88% per decade from
1980–2010). However, the growth rates radically differ from state to state. In some states,
there was even a decline in the number of adult HL speakers.
(2) Spanish and Chinese remain the two most common groups of HLs in the United States,
and their prevalence has grown rapidly over the last 30 years. In addition, although their
numbers are still small in relation to the absolute number of speakers, new HLs such as
Arabic, Hindi, Dravidian, Vietnamese, Russian, Amharic/Ethiopian, and Tibetan, also
experienced substantial growth. At the same time, languages such as French, German,
Italian, Greek, Yiddish, and Dutch are experiencing rapid declines.
(3) Unsurprisingly, states that are typically considered immigration hubs (e.g., California,
Texas, New York, Florida, and New Mexico) have large numbers of HL speakers.
However, in terms of the proportion of HL speakers in the state’s population, some other
states (e.g., Arizona, Hawaii, Massachusetts, Nevada, and New Jersey) have proportions
of HL speakers as large as those in states with high numbers of immigrants.
(4) The majority of HL speakers are speakers of Spanish, but on the state level, there are
quite a few exceptions, such as Alaska (Aleut Eskimo is the most common HL), Hawaii
(Filipino/Tagalog and Japanese), Louisiana (French), Maine (French), New Hampshire
(French), North Dakota (German), South Dakota (Siouan languages), and Vermont
(French).
(5) Over the last 30 years, the concentration of HL speakers has moved from metropolitan
centers to suburbs or non-metropolitan areas. However, some of the HLs continue to have
specific regional concentrations.
At the national level, we observe the transition from the languages former immigrants
groups (e.g., French, German, Italian, Greek, Yiddish, and Dutch) to the languages of recent
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
27!
immigrants, particularly Spanish and Chinese. However, the data also show a sign of incoming
trends in the next few decades. Languages such as Arabic, Hindi, Dravidian, Vietnamese,
Russian, Amharic/Ethiopian, and Tibetan are growing even faster than Spanish and Chinese,
forecasting that, over the next few decades, that the numbers of speakers of these languages may
grow to be as large as that of Spanish or Chinese languages.
At the state or county level, the data reveal a great deal of diversity in the demographics
of adult HL speakers. Each state has a unique composition of HL speakers, and the national
trends do not apply to all the states and counties. For example, the second most commonly
spoken HL at the national level is Chinese, but as illustrated by the compositions of HLs at the
state level, this fact applies to only five out of the fifty total states and one special district (9.8%).
It is a widely acknowledged fact that Spanish is by far the most common HL in the United
States, but, in eight states (15.6%), another language is spoken more frequently than Spanish.
Even among the states where Spanish is spoken most commonly, the proportions of Spanish HL
speakers vary considerably from 38.38% (Pennsylvania) to 76.83% (New Mexico).
It is important that research on HL keep note of such regional differences. In particular,
studies regarding the attrition or maintenance of HLs will lose their generalizability if they do
not take into consideration the influence of the regional demographics of HL speakers where
their data were collected. Using multivariate modeling, Alba et al. (2002) have demonstrated that
the HL community has the most significant influence on the maintenance of HLs (other than
whether their parents are endogamously married). Thus, findings on the attrition or maintenance
of HL in one place (for instance, Philadelphia, where a relatively smaller Spanish community
exists) may not be applicable to another place (e.g., New Mexico, where nearly three-fourths of
HL speakers speak Spanish). Obviously, studies on HL in general should acknowledge such
geographic limitations and should carefully select their data collection sites so that they are
counterbalanced in terms of the demographics of HL speakers.
The density of HL speakers has shown the outward migration of HL speakers in the last
three decades. Ehrenhalt (2013) uses the term “demographic inversion” (p. 3) to describe this
trend of new Hispanic (and black) immigrants such that first- and 1.5-generation immigrants
directly settle in suburbs rather than in central cities. Central cities, which used to offer
manufacturing workplaces and affordable housing for the poor and newly arrived immigrants,
have undergone gentrification and are now attracting wealthy and middle-class residents. As far
as the HL speakers of Spanish are concerned, the outward migration of the affluent from cities to
suburbs has come to an end, and many of them have moved to suburbs much more quickly than
before. On the contrary, some groups of HL speakers, such as Chinese, Filipino/Tagalog,
Korean, Vietnamese, Korean, Dravidian, and Portuguese, still maintain the traditional outward
migration model and reside in the ethnic enclaves with a high density of HL speakers.
This difference between Spanish and other HL speakers, particularly speakers of Chinese
languages, (also see Pew Research Center, 2013) is an interesting parallel with the findings that
Asian (Chinese) HL speakers have a faster attrition rate than Spanish HL speakers (Lopez, 1996;
Alba et al., 2002). Typically, these different speeds of language attrition are attributed to the
typological distance between the dominant language (i.e., English) and the HL (i.e., Spanish is
typologically much closer to English than Chinese), but this study suggests an alternative
interpretation of the higher maintenance rate of Spanish as an HL; that is, Spanish HL speakers’
unique emigration pattern from the metropolitan areas in the recent history. It is an empirical
question whether the higher maintenance of HL is caused by the influence of the demographics
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
28!
of HL communities or the linguistic similarities between the mainstream language and the HL.
It goes without saying that maintenance of HLs involves many other factors, including
HL speakers' language ideologies, degree of cultural assimilation, immigration age, religious
participation in the HL, continuing input in the HL through ethnic media or trips to the country
of origin, motivation, social prestige of the HL, market use/value of the HL, availability of HL
courses/bilingual programs, and isolation or existence of ethnic enclaves, just to name a few. In
recent studies, Lo Bianoco (2014) and Lo Bianco and Payton (2013) argue that creating an
environment to increase vitality and desire to use the HL is critical for maintaining the HL.
It is very likely that there is no single deterministic factor that can predict the likelihood
of HL speakers' maintenance of their HL, since many of these factors overlap with each other
within a complex network of facilitative and hindering relationships. This study has
demonstrated that demographic factors of HL speakers, particularly their absolute numbers and
density, exhibit variation by both region and language. As future research, if one can gather data
for non-demographic variables mentioned above for various language groups in different
geographic locations, it is possible to develop and examine statistical models of these complex
relationships of variables in HL maintenance.
CONCLUSION
Recently, HL speakers have received considerable attention in multi-disciplinary research
in language education, language acquisition, and language policy. This article has presented the
demographics of HL speakers, which are crucial in making decisions about research projects,
allocating resources for HL speakers, and establishing curricula at an educational institution. I
have argued that in addition to the raw number of HL speakers, their composition and density
need to be considered in understanding the demographics of HL speakers in the United States.
These findings of HL speakers’ demographics are also useful for future research on HLs.
This study suggests the possibility that the different attrition rates of HLs between Spanish and
Asian languages may be attributed to the different demographic patterns between these two
language groups rather than to the HLs’ linguistic typological distance from English.
NOTES
1 The!1.5!generation!in!Rumbaut!et!al.!(2005,!2009)!is!defined!as!immigrants!who!
arrive!in!the!United!States!before!the!age!of!12.!Although!the!majority!of!studies!define!the!
1.5-generation!immigrants!as!those!who!arrived!in!the!United!States!during!school!age,!but!
the!exact!cut-off!age!varies!from!study!to!study.!
2 These figures were collected from the ACS data in 2005 and 2007. Unlike this study, in
which the target population is adult bilingual HL speakers, the rankings by Fee et al. (2014) are
based on the numbers of all HL speakers (including young children and first-generation HL
speakers). This different approach to defining HL speakers likely caused these minor differences
in the rankings of HL speakers as of 2005–2007.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
29!
REFERENCES
Alba, R., Logan, J., & Stults, B. (2002). Only English by the third generation? Loss and
preservation of the mother tongue among the grandchildren of contemporary immigrants.
Demography, 39, 467–484.
Au, T-K-F., Knightly, L. M., Jun, S-A., & Oh, J-S. (2002). Overhearing a language during
childhood. Psychological Science, 13, 238–243.
Brecht, R. D. & Ingold, C. W. (2002). Tapping a national resource: Heritage languages in the
United States. Center for Applied Linguistics: Washington D.C.
Carreira, M., Jensen, L., Kagan, O., & Giangrande, M. (2009). The Heritage Language Learner
Survey: Report on the Preliminary Results. Accessed 2 March 2015 at
http://www.international.ucla.edu/media/files/paper.pdf
Carreira, M., & Kagan, O. (2011). The results of the National Heritage Language Survey:
Implications for teaching, curriculum design, and professional development. Foreign
Language Annals, 44, 40–64.
Ehrenhalt, A. (2013). The great inversion and the future of the American city. New York:
Vintage International.
Fee, M., Rhodes, N. C., & Wiley, T. G. (2014). Demographic realities, challenges, and
opportunities. (1). In T. G. Wiley, J. K. Peyton, D. Christian, S. C. K. Moore, & N. Liu
(Eds.), Handbook of heritage, community, and Native American languages in the United
States: Research, policy, and educational practice (pp. 6–18). New York: Routledge.
Fishman, J. A. (2001). 300-plus years of heritage language education in the United States. In J.
K. Peyton, D. A. Ranard, & S. McGinnis (Eds.), Heritage languages in America:
Preserving a national resource (pp. 81–98). Washington, DC: Center for Applied
Linguistics.
Gaquin, D. A. & White–Dunn, G. (2014). The who, what, and where of America: Understanding
American Community Survey (4). Lanham, MD: Bernan Press.
Hornberger, N. H., & Wang, S. C. (2008). Who are our heritage language learners? Identity and
biliteracy in heritage language education in the United States. In D. Brinton, O. Kagan, &
S. Bauckus (Eds.), Heritage language education. A new field emerging (pp. 3–35). New
York: Routledge.
Kagan, O., & Dillon, C. (2003). Heritage speakers' potential for high-level proficiency. In H.
Byrnes & H. H. Maxim (Eds.), Advanced foreign language learning: A challenge to
college programs (pp. 99–112). Boston, MA: Heinle/Thomson.
Lo Bianco, J. & Peyton, J. K. (2013). Vitality of Heritage Languages in the United States: The
Role of Capacity, Opportunity, and Desire. Heritage Language Journal, 10(3).
Lo Bianco, J. (2014). Language vitality and heritage languages. In T. G. Wiley, J. K. Peyton, D.
Christian, S. C. K. Moore, & N. Liu (Eds.), Handbook of Heritage, Community, and Native
American Languages in the United States: Research, Policy, and Educational Practice.
(pp. 54-65). New York, NY: Routledge.
Lopez, D. E. (1996). Language: Diversity and assimilation. In R. Waldinger & M. Bozorgmehr
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
30!
(Eds.), Ethnic Los Angeles (pp. 139–159). New York: Russell Sage Foundation
Publications.
Montrul, S. (2002). Incomplete acquisition and attrition of Spanish tense/aspect distinctions in
adult bilinguals. Bilingual Language and Cognition, 5, 39–68.
Montrul, S. (2009). Knowledge of tense—aspect and mood in Spanish heritage speakers. The
International Journal of Bilingualism, 13, 239–269.
Mori, Y., & Calder, T. M. (2013). Bilingual vocabulary knowledge and arrival age among
Japanese heritage language students at Hoshuukoo. Foreign Language Annals, 46, 290–
310.
Oh, J. S., Knightly, L., & Au, T.–K.–F. (2003). Holding on to childhood language memory.
Cognition, 86, B53–B64.
Pan, S. (2002). Chinese in New York. In O. García & J. Fishman (Eds.), The multilingual apple:
Languages in New York City (pp. 231–255). New York: Mouton de Gruyter.
Pew Research Center, 2013. “The Rise of Asian Americans”. Pew Social Trends:
http://www.pewsocialtrends.org/2013/04/04/asian-groups-in-the-u-s/
Polinsky, M., & Kagan, O. (2007). Heritage languages: In the ‘wild’ and in the classroom.
Language and Linguistics Compass, 1, 368–395.
Polinsky, M. (2011). Reanalysis in adult heritage language: A case for attrition. Studies in
Second Language Acquisition, 33, 305–328.
Potowski, K. (2010). Language diversity in the USA. Cambridge, MA: Cambridge University
Press.
Ruggles, S., & Sobek, M. (1997). Integrated public use microdata series: Version 2.0.
Minneapolis, MN: Historical Census Projects, University of Minnesota.
Rumbaut, R. G., Massey, D. S., & Bean, F. D. (2005). Linguistic life expectancies: Immigrant
language retention in Southern California. Population and Development Review, 32, 447–
460.
Rumbaut, R. G. (2009). A language graveyard? The evolution of language competencies,
Preferences and use among young adult children of immigrants. In T. G. Wiley, J. S. Lee,
& R. Rumberger (Eds.), The education of language minority immigrants in the United
States (pp. 35–71). Bristol, UK: Multilingual Matters.
Ryan, C. (2013). Language use in the United States: 2011. US Census Bureau.
Valdés, G. (2001). Heritage language students: Profiles and possibilities. In J. K. Peyton, D. A.
Ranard, & S. McGinnis (Eds.), Heritage languages in America: Preserving a national
resource (pp. 37–77). Washington, DC: Center for Applied Linguistics.
Wiley, T. G., Peyton, J. K., Christian, D., Moore, S. C. K., & Liu, N. (Eds.) (2014). Handbook of
heritage, community, and Native American languages in the United States: Research,
policy, and educational practice. New York: Routledge.
Zentella, A. C. (1997). Growing up bilingual: Puerto Rican children in New York City. Oxford:
Basil Blackwell.
This is an outdated draft. Visit http://doi.org/10.1111/modl.12272 for the post-print version.
!
31!
Zentella, A. C. (2002). Spanish in New York. In O. García & J. Fishman (Eds.), The multilingual
Apple: Languages in New York City (pp. 167–201). New York: Mouton de Gruyter.
... For the purposes of this study, we identified HL speakers as those who reported using a language other than English at home (whether currently or previously) who were either born in the U.S. or arrived in the U.S. at the age of 13 or younger. This definition has been employed primarily with the practicality of survey administration in mind (see Nagano, 2015). For this survey study, to ensure a high return rate, we had to limit questions identifying HL speakers to two items that we believed most effectively identified respondents as HL speakers. ...
... The proportion of HL speakers in the modern language classroom at community colleges is remarkably high (42%). Using U.S. Census data, Nagano (2015) estimates that approximately 9.2% of the U.S. adult population speaks HLs at home. Accordingly, it is noteworthy that nearly half of language students at community colleges come to the modern language classroom with some degree of proficiency in a language other than English. ...
... No data are available for less commonly taught languages (such as Chinese, Russian, Korean, Tagalog, etc.), but the comparable numbers for these languages must be significantly lower than for Spanish. The unavailability of HL classes is a particularly challenging issue: a recent study on the demographics of HL speakers (Nagano, 2015) shows that HL speakers have increased not only in number (on average a 27% increase per decade in the last three decades) but also in diversity, marked by the rapid growth of HL speakers of less commonly taught languages. Thus, while HL classes are effective and necessary for HL speakers, there are also practical challenges to offering HL classes to all HL speakers, especially those who speak less commonly taught languages. ...
Article
This study analyzes the choices that heritage language (HL) learners make when enrolling in language courses at community colleges. Data from the Students and Instructors of Languages at Community Colleges (SILCC) Surveys, a nationwide survey with 1,756 students taking language courses at 101 community colleges across 33 states in the U.S., show that as many as 42.2% of community college students in modern language classrooms are identified as HL speakers. Surprisingly, more than half of these HL speakers are studying a language other than their own HL despite their prior linguistic knowledge, cultural familiarity, and familial ties with their HL. This paper evaluates a few possible explanations why a large proportion of HL speakers are opting to learn a new, third language. Building upon prior research and current data, we discuss differences in linguistic backgrounds, demographics, motivational attributes, and academic goals between HL learners studying their own HL and those studying a new language.
... For the purposes of this study, we identified HL speakers as those who reported using a language other than English at home (whether currently or previously) who were either born in the U.S. or arrived in the U.S. at the age of 13 or younger. This definition has been employed primarily with the practicality of survey administration in mind (see Nagano, 2015). For this survey study, to ensure a high return rate, we had to limit questions identifying HL speakers to two items that we believed most effectively identified respondents as HL speakers. ...
... The proportion of HL speakers in the modern language classroom at community colleges is remarkably high (42%). Using U.S. Census data, Nagano (2015) estimates that approximately 9.2% of the U.S. adult population speaks HLs at home. Accordingly, it is noteworthy that nearly half of language students at community colleges come to the modern language classroom with some degree of proficiency in a language other than English. ...
... No data are available for less commonly taught languages (such as Chinese, Russian, Korean, Tagalog, etc.), but the comparable numbers for these languages must be significantly lower than for Spanish. The unavailability of HL classes is a particularly challenging issue: a recent study on the demographics of HL speakers (Nagano, 2015) shows that HL speakers have increased not only in number (on average a 27% increase per decade in the last three decades) but also in diversity, marked by the rapid growth of HL speakers of less commonly taught languages. Thus, while HL classes are effective and necessary for HL speakers, there are also practical challenges to offering HL classes to all HL speakers, especially those who speak less commonly taught languages. ...
Conference Paper
This study analyzes the choices that heritage language (HL) learners make when enrolling in language courses at community colleges. Data from the Students and Instructors of Languages at Community Colleges (SILCC) Surveys, a nationwide survey with 1,756 students taking language courses at 101 community colleges across 33 states in the U.S., show that as many as 42.2% of community college students in modern language classrooms are identified as HL speakers. Surprisingly, more than half of these HL speakers are studying a language other than their own HL despite their prior linguistic knowledge, cultural familiarity, and familial ties with their HL. This paper evaluates a few possible explanations why a large proportion of HL speakers are opting to learn a new, third language. Building upon prior research and current data, we discuss differences in linguistic backgrounds, demographics, motivational attributes, and academic goals between HL learners studying their own HL and those studying a new language.
... Questionnaire data on language history and language task performance data previously collected were thus utilized for the present purpose. The two language groups were both included, as these two groups present two of the biggest immigrant HL populations in the USA (Kim & Chao, 2009;King & Ennser-Kananen, 2013;Nagano, 2015), yet differ in many characteristics, including language use, cultural, and socioeconomic aspects, as well as typological distance between their heritage and majority languages (Kim & Chao, 2009;Nagano, 2015;Tao et al., 2015;Tomoschuk, Ferreira, & Gollan, in press). ...
... Questionnaire data on language history and language task performance data previously collected were thus utilized for the present purpose. The two language groups were both included, as these two groups present two of the biggest immigrant HL populations in the USA (Kim & Chao, 2009;King & Ennser-Kananen, 2013;Nagano, 2015), yet differ in many characteristics, including language use, cultural, and socioeconomic aspects, as well as typological distance between their heritage and majority languages (Kim & Chao, 2009;Nagano, 2015;Tao et al., 2015;Tomoschuk, Ferreira, & Gollan, in press). ...
Article
Full-text available
This study investigated the effects of the amount of cumulative heritage language (HL) exposure during three time periods, on heritage and majority language performance in young adulthood, among two distinct groups of immigrant populations in the USA. Within each time period, exposure from three different sources were examined, and amount of cumulative exposure was calculated encompassing exposure from preceding periods. Factors that may modulate exposure effects were also assessed. Results showed that greater cumulative HL exposure from people at home during all three time periods significantly predicted HL skills for both language groups. For effects on English skills, only the Spanish group showed any influences of exposure. These effects were modulated by parental English proficiency. Input from other sources had less impact. The present findings support the role of parental input throughout childhood, adolescence, and early adulthood in improving HL skills, with less noticeable consequences for the majority language.
... With respect to the first point, ecological dynamics can be tracked within and across regions using publicly available data, including census demographic statistics (e.g., Statistics Canada). For example, Nagano (2015) explored the geographic distribution and demographic characteristics of adult heritage language 8 Debra A. Titone and Mehrgol Tiv speakers across the United States based on data from the U.S. Census. Similarly, they found substantial demographic differences in the adult heritage language speakers who resided in distinct regions of the country. ...
Article
Full-text available
In “The Devil's Dictionary”, Bierce (1911) defined language as “The music with which we charm the serpents guarding another's treasure.” This satirical definition reflects a core truth – humans communicate using language to accomplish social goals. In this Keynote, we urge cognitive scientists and neuroscientists to more fully embrace sociolinguistic and sociocultural experiences as part of their theoretical and empirical purview. To this end, we review theoretical antecedents of such approaches, and offer a new framework – the Systems Framework of Bilingualism – that we hope will be useful in this regard. We conclude with new questions to nudge our discipline towards a more nuanced, inclusive, and socially-informed scientific understanding of multilingual experience. We hope to engage a wide array of researchers united under the broad umbrella of multilingualism (e.g., researchers in neurocognition, sociolinguistics, and applied scientists).
... Whatever the reason for the unusual skill ratings, the background profile collected here contrasts with the findings in Mikhaylova (2012) of significant differences in the background of advanced college-level HL and L2 learners. 6 Typically, adult heritage speakers of Russian continue to have exposure to Russian after their arrival in the U.S., and they self-rate their proficiency in Russian rather highly (Nagano, 2015). Fittingly, the prototypical English-dominant college level HL learners in Mikhaylova (2012) self-rated higher in Russian (especially in oral skills) than the prototypical L2 learners, and they also reported a greater frequency of use of Russian, as well as a higher number and wider variety of contexts. ...
Chapter
This study bridges a gap in heritage language (HL) research to date by focusing on learners that do not neatly fit the often-cited definitions and profiles of HL learners (e.g., Polinsky & Kagan, 2007; Carreira & Kagan, 2011). We study the effects of L1 attrition of Russian in teenage HL learners several years after exposure to the HL in the immediate family or community has ceased, due either to international adoption or to other restricted access to Russian speakers. We compare them in the context of re-exposure through a short-term cultural and language immersion program to age-matched foreign language (L2) learners. Each participant in the program produced oral narratives in two sessions, three and a half weeks apart, and the resulting narratives were coded and analyzed for fluency and accuracy. Although the results show significant attrition in the language of the HL learners, they surpassed the L2 learners in several fluency and accuracy measures. We discuss the strategies that both learner groups used in their narratives, as well as implications of these findings for HL acquisition research and pedagogy. https://muse.jhu.edu/chapter/2092628
... A myriad of termsfrom native speakers to quasi-native speakers to bilingual speakers to home-background speakershave been used to refer to HL learners, highlighting the diversity in language proficiency and family relevance. Nagano (2015) summarized four key factors used in research literature to define HL learners according to broad or narrow criteria: 1) ethnic and ancestral background (Fishman and Peyton 2001), 2) language proficiency (Valdés, 2001), 3) identity and sociopolitical circumstances (Hornberger and Wang 2008), and 4) age of arrival, country of birth, and parents' native language (Carreira and Kagan 2011). Scholars working with CHL learners often follow these traditions in defining their subjects, but tend to take an all-inclusive approach and prefer a broad definition of CHL learners. ...
Article
Full-text available
Drawing on data from a learner corpus of blogs, this study explores epistemic expressions used in Chinese heritage language (CHL) writing from a developmental perspective, and aims to provide a better understanding of pragmatic development in heritage language learning context. A total of 6,511 blog entries written by 266 heritage learners from four different proficiency levels were analyzed cross-sectionally. The findings revealed three notable developmental patterns in CHL learners’ use of epistemic markers (EMs): 1) a rapid increase in the frequency and diversity of EMs at the beginning of CHL curriculum, 2) a period of stability from the second quarter onward, and 3) a divergence of frequency and diversity at the advanced level, whereby the diversity of EMs increased again, but the overall frequency of EMs remained unchanged. Significant developmental variability was also found between grammatical sub-groups of epistemic markers. Overall, the study shed light on the development of CHL learners’ pragmatic competence, and demonstrated the effectiveness of learner corpora as a research tool for studies of pragmatics learning.
... In the HL literature, however, more fine-grained definitions of HL speakers have been proposed. For example, Nagano (2015) introduced varying features that the previous studies employed to define HL speakers including students' ethnicity, birthplace, country of birth (and/or age of arrival to the United States), language used at home, proficiency in the HL, and proficiency in English. Fishman (2001) presented a classification of HL speakers based on the sociohistorical perspectives of the United States. ...
Article
This article describes the development process of a project for heritage language speakers of Mandarin Chinese, Spanish, and Japanese at a high-enrollment community college in the northeast United States. This pilot project, funded by the Henry Luce Foundation, aimed to empower minority group students through active reinforcement of students’ heritage language and culture, crossdisciplinary faculty collaboration, and a cohort-based advisement model. The pilot project consisted of mentorship by faculty members who shared the same heritage language and culture, incorporation of heritage language and international business courses into students’ curricula, and weekly seminars that focused on developing students’ heritage culture and professional skills. A cohort of nine students demonstrated that such a model is effective for minority students by achieving a remarkably high completion rate (66.67%) compared with an average rate for public 2-year institutions in the United States (21.20%).
Article
Issue There is a need for greater access to Spanish language services in United States healthcare. One approach to increasing language concordant care is the use of second language skills by healthcare staff. The desire to use second language skills may have unintended consequences when individuals step beyond their language abilities and can cause more harm than do good for limited-English proficiency patients. Medical students are in a unique position that places them at increased risk for inappropriately using second language skills. Evidence: The use of qualified healthcare interpreters has been shown to mitigate some of the disparities seen with limited-English proficiency patients including poorer healthcare outcomes, less access to care, and lower patient satisfaction. In spite of this knowledge, studies have demonstrated the phenomenon of residents and physicians “getting by” without the use of an interpreter, even when they recognized that their language competency was insufficient to provide high quality care. Regardless of language ability, medical students are asked to engage in conversations with Spanish speaking patients that are beyond their level of language competency. Students vary in their perceived language ability and level of comfort engaging in different clinical scenarios with limited-English proficiency patients. Implications: Students are in a unique position of vulnerability to pressures to use second language skills in situations that step beyond their abilities. We explore how hierarchy intensifies previously established factors, including a lack of adequate training or evaluation and other structural barriers, in contributing to medical students’ inappropriate use of Spanish with limited-English proficiency patients. We propose an approach that includes student education, standardization of clinic rules regarding interpretation, and comprehensive faculty development to address this important patient care issue.
Article
Objective This study examines the differing roles of parents in producing bilingualism among second‐generation Asian and Latino Americans, the fastest growing immigrant groups in the United States. Methods We employ the probit model to estimate the likelihood of language maintenance for both ethnic groups using the Integrated Public Use Microdata Series (IPUMS) USA of 2005–2014. Results The estimation results show that mothers play a more significant role than fathers, especially for Latino Americans, and that heritage‐language retention increases with the parents’ age at arrival. We also find an increase in the rates of language maintenance across generations, presumably resulting from heightened awareness of the need to preserve cultural heritage among younger immigrants in recent decades. Conclusion These findings highlight the cultural and structural differences in gendered parenting between the two immigrant groups and suggest potential areas of gains through intervention programs for immigrant parents to promote parental investment in their children's development, including bilingualism.
Article
Full-text available
A framework to examine vitality of languages in a specific context, developed by Francois Grin and elaborated by Joseph Lo Bianco, specifies that three conditions are necessary for language vitality and revitalization: Capacity Development, Opportunity Creation, and Desire (COD). This framework was developed as a tool to help communities and governments support regional and minority languages and to promote policy development at the national level related to language revitalization and use. The framework is used in this issue as a guide for examining the vitality of languages spoken in the United States as “heritage” languages, which are spoken by individuals who have home, community, and intergenerational connections with the languages as well as some proficiency in them.
Article
In most post-secondary Spanish language programs in the U.S., heritage language (HL) learners and second-language (L2) learners are enrolled together, in the same courses (Ingold, Rivers, Tesser, & Ashby, 2002). Nevertheless, there is scant empirical research on what actually goes on in these classrooms and what the nature of learner-learner interactions is (Blake & Zyzik, 2003; Bowles, in press). This study reflects a reality in many Spanish classes because it analyzes the task-based interactions of nine learner-learner pairs, each containing an L2 learner and an HL learner of Spanish. The pairs completed a series of three tasks — one oral (a spot-the-differences task) and two written (a crossword puzzle task and a cloze/complete-the-story task). Results showed that, in completing the tasks, the L2 and HL learners had their language-related issues resolved in equal proportion but that there were qualitative differences on the written tasks in terms of the linguistic targets. HL learners relied on their L2 partners for orthography issues (spelling and accent placement), whereas L2 learners relied on their HL partners for vocabulary issues and, to a lesser extent, for grammar-related queries. Implications for pedagogy in classrooms enrolling both L2 and HL learners are discussed.
Article
What are the most widely spoken non-English languages in the USA? How did they reach the USA? Who speaks them, to whom, and for what purposes? What changes do these languages undergo as they come into contact with English? This book investigates the linguistic diversity of the USA by profiling the twelve most commonly used languages other than English. Each chapter paints a portrait of the history, current demographics, community characteristics, economic status, and language maintenance of each language group, and looks ahead to the future of each language. The book challenges myths about the ‘official' language of the USA, explores the degree to which today's immigrants are learning English and assimilating into the mainstream, and discusses the relationship between linguistic diversity and national unity. Written in a coherent and structured style, Language Diversity in the USA is essential reading for advanced students and researchers in sociolinguistics, bilingualism, and education.
Article
This study examines bilingual vocabulary knowledge in relation to arrival age among first language (L1) Japanese students attending hoshuukoo (i.e., supplementary academic schools for Japanese-speaking children) in the United States. It also examines the relationship between L1 Japanese and English as a second language (L2), as motivated by Cummins's (1979, 1991) notion of linguistic interdependence. One hundred and twenty-two high school students ages 15–18 from eight hoshuukoo took Japanese and English vocabulary tests designed by Ono (1989). Students who came to the United States by age 9 or younger were three grades behind in L1 Japanese and were either ahead of or at their U.S. grade level in English. In contrast, those who arrived at age 10 or older were just one grade behind in Japanese and were two to five years behind in English. High vocabulary knowledge in one language was associated with low knowledge in the other, and the negative correlation between L1 and L2 became statistically nonsignificant when arrival age was controlled. Consequently, arrival age remains an important factor in accounting for hoshuukoo students' bilingual vocabulary learning, and the notion of linguistic interdependence must be reexamined in factors in addition to vocabulary knowledge.
Article
Conversational interaction studies have typically focused either on second language (L2) learners participating in native speaker–nonnative speaker (NS–NNS) dyads or in NNS–NNS dyads. This study analyzes the task-based interactions of 26 naturally occurring learner dyads in an intermediate-level, university Spanish language classroom, 13 of which were matched L2 learner dyads and 13 of which were mixed L2 learner–heritage learner (HL) dyads. Specifically, the study compared the two dyad types to determine whether they differed in their focus on form or in the amount of talk produced during interaction. Results revealed that the two types of dyads were largely similar, although instances of focus on form were more likely to be resolved in a target-like way by mixed L2–HL pairs than by matched L2–L2 pairs, and there was significantly more target language talk in mixed pairs. Interestingly, L2 learners used the target language significantly more with HL learners than they did with other L2 learners, suggesting that different conversational norms may be at play in the two pair types. Furthermore, posttask questionnaire data indicated that L2 and HL learners alike saw the interaction as a greater opportunity for the L2 learner's development than for the HL learner's, calling into question whether classroom contexts like this one meet the needs of HL learners.