English Spoken Here?

What the 2000 Census Tells Us about Language in the USA

by Dennis Baron


The 2000 US Census reports a 47% increase in speakers of languages other than English in the United States during the 1990s, compounding a 38% rise in the 1980s. In addition, the Census reports more people speaking a language other than English at home. Such figures stimulate nativist reactions against immigration and generate calls for making English official. Last year, Samuel P. Huntington blamed what he saw as the decline of English on Hispanics, and he warned that unless America returns to its Protestant, anglophone roots, it will split into two nations. The Spanish-speaking population of Iowa doubled in the decade of the 1990s to 2.9%. While that figure represents only 80,000 people, most of whom also speak English, the state's voters saw fit to make English Iowa's official language. More recently, the governor of West Virginia vetoed an official English bill on a technicality.

A look at what the 2000 Census tells us about language use in the United States reveals that English continues to be the dominant, and in many cases, the only language of the United States and confirms that warnings about the decline of English are unwarranted. As they have done in the past, non-English-speakers almost inevitably switch to English, and many of them are doing so much more quickly than past generations have done.

fig. 1 Non-English languages spoken in the U.S.
[Language Use and English-Speaking Ability: 2000, p. 8]

The 2000 Census reports a 47% increase in speakers of languages other than English in the United States during the 1990s, compounding a 38% rise in the 1980s. This reflects a jump in immigration, but it’s the kind of headline-grabbing statistic that causes Americans to fear that English, the language that has taken over the globe, is actually in danger at home (see fig. 1).

Not to worry: English is secure. The number of immigrants acquiring English closely tracks the rise in immigration, so despite the concerns of English-only advocates, there has been no net loss in English usage in the U.S. The real endangered languages in America continue to be the ones spoken before the English came, and the ones that immigrants bring with them. Families tell stories of grandmothers who never learned English and lived out their lives in Italian, Polish, Chinese or Spanish neighborhoods, where all their needs were met in their native language, but the fact is that even immigrants who try to hold on to the old ways will lose their language. Of the 47 million minority-language speakers over five years old now in the country, 43.6 million of them speak at least some English, and over half of them speak English fluently. With time, the rest will achieve fluency or something close to it. And their descendants will likely become monolingual English speakers.

Repeating patterns established during high periods of immigration in the late nineteenth and early twentieth century, today’s newly-immigrated non-English-speakers often find work in low-status, low-paying jobs. Middle class life, if it comes, may wait a generation or two, by which time English is also well-established. Nonetheless, the perception remains that nonanglophones are taking over the country, and we regularly hear nativist calls for tighter immigration controls, an end to bilingual education, and the adoption of English as the official language of the United States and of the several states. Echoing fears voiced by nativists a century ago, today’s nativists also want to restrict both the public and private scope of other tongues.

Critics in earlier generations targeted German as the greatest threat both to English and to national stability. Benjamin Franklin and other 18th-century nativists accused German Americans of taking jobs away from English workers, of speaking a debased dialect of their own language, and of refusing to learn English. Franklin considered Germans to be racial “others,” calling them “swarthy” in contrast to the pale-faced English.  Other Anglos accused the Pennsylvania Germans of laziness, illiteracy, clannishness, a reluctance to assimilate, excessive fertility, and Catholicism. They were even blamed for Pennsylvania’s harsh winters. Early language laws requiring English in the schools, courts, and government offices, on licensing exams, in newspaper ads and on business signs, were really attacks against immigrants rather than attempts to protect the English language in the United States.

fig. 2  Density of speakers of languages other than English in West Virginia
[MLA Language Map www.mla.org]

Today’s nativists see as their enemy not the Germans but the growing number of Spanish speakers. The new nativists argue that unlike previous immigrant groups, Hispanics show more language loyalty and resist assimilation into the Anglo mainstream.  And they’re taking action before the situation gets out of hand. According to the 2000 Census, West Virginia has the highest percentage of English-only speakers in the U.S. (see fig. 2). But with only 2.5% of the population of West Virginia speaking languages other than English, two weeks ago legislators passed a Parks and Recreation bill with a rider making English the official language of the state.  Responding to a clamor in the press, many legislators claimed they had no idea that they had voted for an English-only law. The governor, a long-time supporter of official-English legislation, may have sensed that sneaking a language law onto the books was not good for his political image, and he vetoed the bill on a technicality: West Virginia’s constitution requires that a law treat only one subject.

Other heavily English-speaking states have been more decisive in passing official language laws to guard English against the imaginary menace posed by speakers of other languages. In 2002 Iowa’s voters, panicked over the doubling of Spanish speakers in the state to 2.9% of the over-5 population, made Iowa the twenty-seventh state to go English-only, even though most of Iowa’s 80,000 Hispanics speak English as well as Spanish.

fig. 3. In Postville Iowa, pop. 2,273, immigrant neighborhoods are sharply delineated: 24.7% of the residents living north of 100th St. speak a language other than English at home; the figure for residents south of 100th St. is 5.3%.


Here’s another way to look at the linguistic dangers Iowa’s anglophones face. It’s true that in tiny Postville, Iowa, population 2,273, about a fifth of the residents speak Spanish, and just over 3% of the townspeople speak Yiddish (see fig. 3). In New York City, with a population of about 7.5 million, just under 25% of the residents speak Spanish, while only 1% speak Yiddish. Although Hispanics have doubled in Iowa, the percentage of people speaking Yiddish in Postville has skyrocketed, more than tripling in recent years. Just the opposite has happened in New York, where the percentage of people speaking Yiddish has declined drastically since the early 1900s. Person for person, three times as many Postvillians speak Yiddish as New Yorkers.  English in Postville, which boasts the largest and until recently the most controversial kosher meatpacking plant in the world, may actually be more “threatened” by the incursion of Yiddishkeit than English in New York City. As for Spanish, it doesn’t seem to pose much of a threat to English: 83% of the Hispanics in Postville, and 88% of New York’s Hispanics, report speaking English “very well.”

Hispanic language loyalty

In the traditional American immigrant paradigm, newcomers speak their native language; the second generation is bilingual; and the third generation speaks only English. But in his most recent book, Who Are We? The Challenges to America’s National Identity, Harvard’s Samuel P. Huntington (2004) insists that this model no longer applies. Huntington, who in an earlier polemic, The Clash of Civilizations, alerted Americans to the Arab threat to life as we know it, now complains that Hispanics are departing from earlier immigrant patterns of linguistic assimilation. Huntington charges that American Hispanics retain Spanish and at best become bilingual, rather than switching completely to English, and he warns that unless America goes back to its Anglo-Protestant, anglophone roots, it will split into two nations, one speaking English, the other, Spanish.

Claims by Huntington and the other new nativists are contradicted by language data from the 2000 U.S. Census and the follow-up “2002 Survey of Latinos” conducted jointly by the Pew Hispanic Center and the Kaiser Family Foundation. These surveys reveal both a high degree of assimilation among American Hispanics, and no real difference between Hispanics and other immigrant groups in the ability to use English.

That is not to say that Hispanics are behaving exactly like other immigrant groups (and while most Hispanics in the U.S. today are immigrants, it’s also important to note that many Hispanic families lived in the Southwest long before the Anglos came). Although the Pew survey confirms earlier observations that Hispanics place a high value on adopting English, there are some significant indicators of Spanish language loyalty as well. Bilingualism persists even in the third generation in border areas where Mexican immigration and the need to interact with Spanish-speaking newcomers remain high. Dominicans also exhibit a high degree of bilingualism, due in part to the fact many migrate back and forth between the U.S. and their ethnic home.

The Pew study notes that third-generation Latinos who exhibit more bilingualism than other immigrant groups are almost all descended from immigrants who came to the United States before the immigration boom of the 1970s (Pew 2004b, p. 3). 50% of these third-generation Latino adults report fairly proficient Spanish alongside their English (2004b, 5), though only 10% of bilinguals from any generation use only Spanish, or more Spanish than English, on the job, and a third speak either only English, or more English than Spanish, at home, the place where one might expect heritage languages to flourish (Pew 2004b, pp. 2-3).

In any case, third-generation bilingualism is nothing new. Both German and Canadian French persisted among U.S. immigrants across several generations as well (Alba, 9). What is important to note is that not all Hispanics keep their heritage language, and there is countervailing evidence that Hispanic assimilation can be so rapid that Spanish starts to disappear as early as the second generation. 14% of second-generation Hispanics report that they speak only English, while a mere 6% are Spanish-only speakers. 80% of second-generation Hispanics are bilingual, with 92% of them speaking English well or very well, and half of them indicating that they are actually English-dominant (Pew 2004a, 2004b; Alba 2004). And third-generation bilingualism among Hispanic children may be a temporary condition. Since many of these children also exhibit a relative lack of Spanish fluency, they could easily become monolingual English speakers as adults.

Although all immigrant groups show clear evidence that they are assimilating, Huntington insists that the newest Hispanic arrivals will hold on to their language longer than older immigrant populations did. But Alba reports that only 7.7% of second-generation children, and only 2.9% in the third generation, don’t speak English well (Appendix, table 1), and he adds that among Mexicans, the largest immigrant stream in the 1990s, assimilation to English by children not only did not weaken, it may have actually increased (7). The Pew study confirms that the largest numbers of English-dominant and English-only speakers among Hispanics turn out to be the youngest as well (Pew 2004b, p.5). While bilingualism persists among pre-1970s immigrants, post-1970s Hispanic immigrants may not hold on to their heritage language for long at all, and first-language loss may even be accelerated for these latest newcomers.

The Census language question

The Census language question is designed to give a snapshot of linguistic conditions in the United States during the period in which the Census was administered. There is only one three-part language question on the Census, and while analyzing responses to that question can be fruitful, it may be unwise to assess the assimilation of today’s immigrants based solely on their responses to this question.

       The 2000 U. S. Census asks respondents whether a language besides English is spoken in their homes. If another language is used, they are asked to identify it and to indicate whether any household members also speak English, and if so, how well.

11a. Does this person speak a language other than English at home?

c    Yes

c    No -> Skip to 12

b. What is this language?

c c c c c c c c c c c

(For example: Korean, Italian, Spanish, Vietnamese)

c. How well does this person speak English?

c   Very well

c    Well

c   Not well

c   Not at all

fig. 4. The language question from Census 2000

The Census reveals that by 1970, Spanish had replaced German as the most commonly-spoken non-English language in the United States. Other languages moved up in the rankings, tracking immigration trends: Chinese rose from 5th place in 1990 to second in 2000; Vietnamese jumped from 9th to 6th place, Russian from 15th to 9th and Arabic from 13th to 11th. Korean remained in 8th place, while French dropped from 2 to 3, German from 3 to 4, Italian from 4 to 7, Japanese from 11 to 13, and Hindi from 14 to 16.

But these rankings may not reflect the actual distribution of languages spoken in the U. S. because the answers that respondents give to question 11a. may not be altogether reliable. Joshua Fishman (1985) has shown that sometimes the pressure to assimilate causes under-reporting of foreign-language use; but at times when ethnic difference becomes trendy, other-language use may actually be inflated.

Linguistic isolation

In addition to language demographics, the Census tracks the prevalence of what it calls “linguistic isolation,” a factor which highlights language loyalty, the degree to which speakers continue to speak their heritage language in the face of competition from English. Census Bureau analysts consider households linguistically isolated if either

a.     no teen-ager or adult speaks only English or

b.     if no one checks the box labeled “speaks English very well.” Such households, it is presumed, will hold on to their native language rather than learning English.

According to these criteria, 24% of Hispanic households are linguistically isolated, while only 15% of households where other Indo-European languages are spoken – including Hindi and Urdu – get that label. With an isolation rate of 29%, Asian-language households in the United States are the least assimilated linguistically at the moment the Census snapshot was taken. That picture is likely to change: as Dorothy Waggoner (1988) showed in her analysis of earlier Census reports, the longer that immigrants stay in the U.S., the better their English becomes. In addition, the 2000 Census indicates that American-born children of Asian immigrants tend to have the highest percentage of monolingual English speakers, and 72% of third-generation or later Hispanic children speak English exclusively.

Even if one lives in a linguistically-isolated household, that doesn’t preclude future language development. While it’s possible that English will gain no purchase there in the years to come, it’s much more likely that the Census snapshot actually reflects a family in transition from first- to second-language fluency, whose members in ten or twenty years will either be comfortably bilingual, or they will have become monolingual English speakers.

English fluency

Question 11c. on the Census asks respondents who speak a language other than English to rate how well they speak English. The choices are “very well,” “well,” “not well,” or “not at all.” According to the Census analysis of answers to this question, foreign-born speakers of Spanish currently report weaker English than foreign-born speakers of Indo-European or Asian and Pacific Island languages. 28% of Spanish speakers speak English “very well,” while 42% of Asian and Pacific Island language speakers and 55% of Indo-European language speakers identify as fluent in English. In contrast, native-born speakers of Spanish report levels of English comparable to native-born speakers of Asian languages, and both groups lag only slightly behind speakers of Indo-European languages. This suggests that while Hispanic immigration is proportionally higher than that of other groups, second-generation Spanish speakers are acquiring English with little difficulty, a finding that is confirmed by the Pew study.

Using U.S. Department of Education guidelines, the Census groups everyone reporting a level of fluency below “speaks English very well” into the broad category “speaks English with some difficulty.” This includes those who claim to speak English “well,” “not well,” and “not at all.” If we combine monolingual English speakers with other-language speakers who speak English “very well,” then 91.9% of Americans have no difficulty with English. In 1990, 93.9% of the population reported very fluent English, and in 1980, 95.2% of all Americans, whether or not they also spoke another language, claimed to speak English very well.

The new nativists look at this fluency data and conclude that there has been a 3.3% decline in English-speaking ability over twenty years, further evidence, to them, that aliens are displacing English at a dangerous pace. But the Census fluency data is particularly unreliable, because respondents have no objective standard for deciding whether they speak English “well” or “very well,” and they may also disagree about what constitutes speaking English “not well” or even “not at all.” Census takers are not competent to make these distinctions either. It would be dangerous to conclude from the answers to question 11c. that English fluency in American is actually getting worse.

There’s a further problem with the language competency data. The Census only asks speakers of languages other than English about their English fluency. If residents of households where only English was spoken were asked to rate their proficiency, the results might be surprising. When I ask my monolingual English students, all English majors, to rate their ability to use English, almost no one chooses the highest category. Students for whom English is not their first language – also English majors – rate themselves lower still. To an outside observer however, both groups of English majors exhibit native-speaker competence.

I’m not claiming that users of other languages are simply being modest when they answer the Census question about English fluency. What I am saying is that their answers cannot be plotted against a standard scale since no such scale exists. It would be safer to read the Census data on competence as a measure of linguistic insecurity – the degree to which speakers lack confidence in their use of English – rather than a picture of actual fluency patterns.

What is to be done?

Whether or not fluency is down, immigration is up and there is definitely a need to provide more transitional language services for nonanglophones. But it has been the case from the 1900s on that the demand for English-language instruction has outstripped the availability of classes in English as a Second Language (ESL), leaving newcomers to pick up English when and where they can. While the percent of American residents who don’t speak any English doubled between 1980 and 2000, it’s still only 1.3% of the 262,375,152 Americans over five years old. But that small percentage translates to three million nonanglophones, in addition to the close to seven million who say they speak English, though not very well. Ten million people whose English is iffy or nonexistent is a significant number. We can’t just tell them, get English or get out.

Immigrants can’t possibly master English overnight, even though the new nativists wish that they would. Learning a language takes more than motive and opportunity, it takes persistence. Adults who manage to find space in an ESL class often discover that the pressures of work, family, or simply traveling to class make regular attendance difficult. American schools, never very effective at teaching native English speakers a foreign language, have also never done well teaching English to immigrant children. In the early twentieth century, at the height of the pre-World War I immigration boom, few schools even tried to deal with the influx of nonanglophones into English-only classrooms, and consequently dropout rates soared.

The 1974 Supreme Court decision in Lau v. Nichols demonstrated that American schools were still immersing and ignoring their non-English-speakers. Both the bilingual programs that flowed from Lau and the new immersion programs that arose in response to complaints about bilingual education have proved unpopular and ineffective. Children still pick up more of their English from other children and from the media than from instruction in the classroom. And they still become fluent in English, though not necessarily in the English of the schoolroom.

This fluency in spoken English serves newcomers well. According to the Pew study, which goes beyond the Census to survey ability to read and write alongside spoken language competence, success in speaking English is still enough to carry immigrants into the American middle class, even if it’s not accompanied by proficient reading and writing. The Pew survey also informs us that bilingual Hispanics use English, not Spanish, on the job, and that they prefer to read job-related materials such as manuals and instruction sheets in English, not Spanish.

But while the acquisition of English can lead to economic benefits, it doesn’t necessarily reduce discrimination on or off the job. Although Spanish-dominant Latinos claim that they are treated unfairly because of their language, those who are English-dominant continue to report discrimination, attributing it not to language but to physical appearance (Pew 2004c, 80). The resentment of foreign languages shown by advocates of official English legislation is just a stand-in for a more deep-seated xenophobia.

From the outset, American history has been a history of immigration, both voluntary and involuntary. It has been a history of the clash of civilizations, of the merging and submerging of cultures and languages. So far, in the clash of languages in the United States, English has come out on top. It’s the 800-pound gorilla that has turned America into “a veritable cemetery of foreign languages” (Portes and Hao, cited in Pew 2004b, 3). Having made its mark at home, English now threatens to become master of the universe as well.

But as a multicultural nation, the United States has always dealt with speakers of many languages within its borders. The only time when this was not the case was the period between World War I and the 1960s, the English-only age that the new nativists would like to return to. After the shock of World War I, nativist-driven reform shut immigration down, and with no large influxes of new populations, those already in the U.S. who hadn’t already assimilated did so, with almost everyone adopting English. The 1965 Immigration Reform Act reopened the nation’s doors to new waves of immigrants from new regions, and America had to get used to foreign languages on its streets and in its classrooms once again. Once again it responded partly with welcome, partly with alarm. Once again we hear calls to close American borders and restore English as the language on everyone’s lips. California’s immigrant governor Arnold Schwarzenegger spoke for many xenophobic Americans in April, 2005, when he supported the idea of private posses working the southern border to keep out illegal immigrants.

But the most recent data from the 2000 Census and the 2002 Latino Survey shows that the adoption of English by non-English speakers continues at impressive rates and that, while bilingualism persists to an extent that worries nativists, English is still the language on everyone’s lips, not to mention their t-shirts, their iPods, and their televisions.

Which is not to say that American language policy should either ignore or condemn the many residents who speak languages other than English.  For language policy in the United States to be effective, it must flow not from ideologies of nativism and protectionism, but from social and linguistic realities. Those realities reveal nonanglophones with an abundance of motivation who continue to switch to English at rates which actually alarm their own ethnic communities. Despite the popularity of foreign-language cable television, radio, and newspapers, immigrants switch to English-language media as they themselves become English-dominant. Another staple of ethnic communities, afternoon or week-end schools to maintain heritage language and culture for immigrants’ children, doesn’t work well either to preserve language: most of the children would rather be anywhere else. By the time they get to college, these same students, South Asian, Hispanic, Korean, Arab or Greek, lament the loss of their heritage language and take steps to try to recover what they can before it’s too late. Unfortunately, even if they master the cuisine or the dance, most of these young adults can’t recover the language sufficiently to converse with relatives when they visit the old country, let alone aspire to any sort of fluency.

What little American language policy there is tends to encourage such language loss. Several scholars have noted the irony that the U. S. first strips newcomers of their native language, then forces them to learn a foreign language that they will never become fluent in, in school. A government expert once remarked to a congressional committee that this wasn’t really a problem, since the languages immigrants lose are neither culturally significant nor vital to the national interest. Rather than protecting an English language that can take care of itself, or worrying about an American Babel that is not to be, our policy makers might do better to address the massive loss of language that the inevitability of assimilation still guarantees. In the “global” 21st century, language is an endangered resource that we cannot afford to squander.

Works cited

Alba, Richard. 2004. Language assimilation today: Bilingualism persists more than in the past, but English still dominates. Albany, NY: Lewis Mumford Center for Comparative Urban and Regional Research, University at Albany.

Census 2000. 2003. “Language Use and English Speaking Ability: 2000.” Washington, D.C.: U.S. Census Bureau C2KBR-29, October.

Fishman, Joshua, et al. 1985. The rise and fall of the ethnic revival. Berlin: Mouton.

Huntington, Samuel P. 2004. Who are we? The challenges to America’s national identity. New York: Simon and Schuster.

“Language Use and English-Speaking Ability: 2000.” 2003. U. S. Census Bureau. www.census.gov.

Pew Hispanic Center. 2004a. Assimilation and language. Washington, D.C.: <pewhispanic.org/site/docs>

_______________. 2004b. Bilingualism. Washington, DC: <pewhispanic.org/site/docs>

_______________. 2994c, 2002 National survey of Latinos. Washington, D.C.: <pewhispanic.org/site/docs>

Waggoner, Dorothy. 1988. Language minorities in the United States in the 1980s: The evidence from the 1980 Census. In Sandra Lee McKay and Sau-Ling Cynthia Wong, Language diversity: Problem or resource? New York: Newbury House. Pp. 69-108.