Ethnicity, Language and Immigration Thematic Series
Interpreting and presenting census language data

Acknowledgements

This paper was authored by Jean-François Lepage, Senior Analyst with Statistics Canada’s Centre for Ethnocultural, Language and Immigration Statistics (CELIS).

The author would like to thank all the internal and external reviewers for their insightful comments that helped improve this paper in many ways. Special thanks go to Jean-Pierre Corbeil, Assistant Director, Diversity and Sociocultural Statistics and in charge of CELIS, and to Eric Caron Malenfant, Chief in CELIS, for their close involvement in its production. The author would also like to thank all his colleagues at Statistics Canada who contributed directly or indirectly to producing, checking, translating and disseminating this paper.

Introduction

Canada outranks other countries around the world for the number of language questions in its Census of Population. No other national census has as many questions about language as the Census of Canada. In 2016, seven questions provided information on mother tongue, knowledge of languages, and languages spoken at home or used at work by Canadians. The language information derived from the Canadian Census of Population is rich, but it is also complex. Canadians declare several hundred languages on the census, and an increasing number of respondents report more than one language in their responses to the different language questions. Moreover, language questions can be combined to create derived variables, indexes or other indicators, which makes data processing and presentation more complex.

Two broad approaches can be used to determine how statistical data on language derived from the Canadian Census of Population are prepared and disseminated. The first involves language groups, while the second concerns the languages themselves as well as the population’s language practices and characteristics. These two approaches are not necessarily incompatible or contradictory; they are simply based on different choices for presenting the data. These choices are closely related to the underlying analytical perspectives of each approach.

The language group-based approach can be dubbed the “standard” approach. For many years, it has been used to process and present Statistics Canada’s census language data. It addresses demolinguistic concerns, a branch of demography that has grown significantly in Canada since the 1960s, mainly as a result of works by the Royal Commission on Bilingualism and Biculturalism (Martel and Pâquet 2010). The root of the demolinguistic problem is the comparison of demographic characteristics or behaviours and, by extension, sociocultural, economic and other characteristics and behaviours, between language groups (Maheu 1985). In Canada, these comparisons mostly concern anglophones and francophones. In an attempt to produce an exhaustive distribution of the population in mutually exclusive groups, a residual category called “allophone”^Note was created. It can be included or excluded from the analysis based on context-specific needs and objectives.

Where language statistics are concerned, the approach based on languages and language practices and characteristics can be called the “emerging” approach. The convergence of three major trends in Canada in recent decades exposed certain limitations in the standard approach and warranted the need for new perspectives. First, steady and more and more linguistically diverse immigration since the 1980s has led to significant growth in the size and heterogeneity of the “allophone” group. This diversification in Canada’s “linguistic landscape” generated growth in multilingualism (or speaking more than one language at home or at work), resulting in increasingly porous borders between language groups. Finally, increasing recognition of the importance of Indigenous languages in Canada, which led to the Indigenous Languages Act being adopted recently by the federal government, stimulated interest in statistics on languages other than English or French or on language practices and characteristics of populations that are not necessarily defined by these language criteria.

Traditionally, the language group approach was used because it generally addressed the concerns expressed by the main social and political stakeholders.^Note In the last decade, Statistics Canada has observed new trends and limitations in the standard approach when it comes to depicting the richness, complexity and diversity of language situations and behaviours in Canada. The agency therefore implemented measures to meet the emerging needs in data and in statistical analyses on language issues in Canadian society.^Note One important measure was to gradually introduce elements that put the focus on languages—or on one language in particular—when preparing, presenting and analyzing data. Meanwhile, we continued to make statistical information on major language groups available. The purpose was to further leverage the information available while maintaining the existing language data offer.

It is important for data users to understand what guides Statistics Canada’s choices in collecting, compiling, analyzing, abstracting and disseminating statistical information on languages in Canada. The agency publishes many data products and analytical products on language issues and, although one of its responsibilities is to consider emerging needs and issues, it is just as important for Statistics Canada to ensure that it does not hinder the traditional, established uses of its data, which remain fully legitimate and relevant.

The aim of this paper is to describe the two main approaches and explain how each requires different choices in organizing and presenting language data for dissemination. Statistics Canada has produced this document to provide language data users with key information on what data are available so they can know which data sources can meet their needs.

The first part of the document presents the language questions asked in the 2016 Census, as well as how the data derived from these questions were presented and made available in Statistics Canada’s different data products, including the agency’s website. The second part provides a more detailed look at the language group approach and the one based on the population’s languages, practices and characteristics. It explains how data are used to produce analyses that answer the specific questions raised by both approaches.

PART 1 — Census language questions and classifications

In the 2016 Census, Statistics Canada collected information on different language characteristics of Canadians using five questions, including two two-part questions.^Note

Most questions (knowledge of official languages, languages spoken at home, first language learned in childhood and still understood) were both in the census short-form and long-form questionnaires^Note (100% data, see Table 1). They were asked to the entire Canadian population.



The other language questions (knowledge of non-official languages, languages used at work) were asked only in the long-form questionnaire (sample data). For the 2016 Census, 25% of households received the long-form questionnaire.

Appendix A presents the census questions as they appeared in the 2016 Census.

The wording of the language questions has been practically unchanged since 2001.^Note The main changes introduced during this period were the addition of instructions in the electronic form to ask respondents who report certain languages (such as “Chinese”) to specify which language they are referring to (Mandarin, Cantonese, etc.), and increasing the number of languages for which data are disseminated.

1.1 Mother tongue, languages spoken at home and languages used at work

The census language questions are used to produce “direct variables”: mother tongue, knowledge of official languages, language spoken most often at home, other languages spoken at home on a regular basis, etc. Data on mother tongue, language spoken most often at home and other languages spoken at home on a regular basis are usually presented in a broad classification (see list below) or a detailed classification which, in 2016, included as many as 269 categories^Note (see all categories in Appendix B).

Below is the broad list^Note for mother tongue, language spoken at home and language of work in the 2016 Census:

• Single responses
  • English
  • French
  • Non-official languages
    • Indigenous languages
    • Other non-official languages
• Multiple responses
  • English and French
  • English and non-official language
  • French and non-official language
  • English, French and non-official language

The categories that make up these classifications do not represent only distinct languages. They also include language families (Inuit languages, Indo-European languages, Germanic languages, etc.), main combinations (single responses, official languages, Indigenous languages, etc.) and residual categories (“not included elsewhere” or “not otherwise specified”). The terms used to designate the different categories are generally stable, but may vary in some analytical documents (see Figure 1).

Description for figure 1

• Official languages
  • English
  • French
• Non-official languages (other languages¹)
  • Indigenous languages
  • Non-Indigenous languages
    • Immigrant languages¹
    • Sign languages

1. These terms are used in analytical documents, but are not found in classifications.

The number of categories in the detailed classification increases from one census to the next. In 2001, the detailed classification comprised 160 categories. This rose to 186 categories in 2006, and to 232 categories in 2011. In order for data on a specific language to be disseminated, they must meet certain quality criteria and maintain confidentiality of responses. The number of languages that meet these criteria increases constantly from one census to the next.

1.2 Knowledge of languages and first official language spoken

The classifications used to disseminate data on knowledge of official languages and on the derived variable “first official language spoken” (FOLS) are much shorter because they exclusively concern Canada’s official languages, English and French (see Table 2).



Data on knowledge of non-official languages are generally presented together with official languages data. A detailed classification identical to the one used for mother tongue data (see Appendix B) is used, excluding multiple response categories. Unlike data on mother tongue, language spoken at home or language used at work, multiple responses are not usually presented separately for knowledge of non-official languages. For each language enumerated, the data on language knowledge provide the total number of people who speak that language, i.e., the number of respondents who reported being able to conduct a conversation in that language, whether they are unilingual, bilingual or multilingual. The “English” category therefore indicates the number of respondents who reported being able to conduct a conversation in English, the “French” category the number of respondents who reported being able to conduct a conversation in French, and so on. Respondents who are able to conduct a conversation in English and in French will be enumerated in both the “English” and “French” categories. With this data presentation method, the sum of the categories is greater than the total Canadian population.

That said, two derived variables provide further details on this. Data on language knowledge are sometimes cross-tabulated with a variable called “knowledge of languages: single and multiple language responses”^Note (see Table 3). For each language, this variable is used to count the number of respondents who reported being able to conduct a conversation only in this language (unilingual) and those who also reported being able to conduct a conversation in at least one other language (bilingual or multilingual). However, it is not possible to identify the specific combination of languages, if any, when data are presented this way.



This gap is partly filled by another derived variable, the “number of languages known,” which presents certain combinations of languages known (Table 4). However, these combinations do not identify a particular language other than English or French.



PART 2 — Main approaches

The different ways to present census language data can be divided into two broad approaches. The main difference between them is how they process multiple responses (or combinations of responses).^Note This distinction is nonetheless based on a significant difference in perspective.

The first approach looks at groups, defined by language characteristics, which make up a population. It attempts to understand and document relationships between language groups in demographic terms (relative population growth, pull, gains and losses in exchanges with other groups, etc.).^Note Reduced to the essential, this perspective is mostly interested in the number of people who make up each group, in the respective demographic evolution of the groups, and in the factors that determine this evolution.

The second approach looks at language dynamics, characteristics and behaviours within a population. It explores linguistic diversity, the knowledge and use of certain languages, bilingualism and multilingualism. It seeks to understand language experience, the relationship with languages, language learning and retention, language vitality factors, etc.

This section of the paper presents these approaches in greater detail, and illustrates the relationships between them and how language data are used in each.

2.1 Language group approach

Census data can be used to define a population or subpopulation of interest according to specified criteria. Users can get statistics on different groups, including Indigenous peoples, youth, women or immigrants, for research, program or policy development, or for any other reason. Language data from the Census of Population can also be used to define a population by its language characteristics. Users can, for example, identify a “francophone” or an “anglophone” population based on a criterion or a combination of selected criteria, and thereby examine a specific language group, or they can divide the population of a country, province or region into different language groups.

The objective of the language group approach is to study relationships between the groups in a population. By extension, it also explores the characteristics and behaviours of a language group, generally for comparison purposes. These characteristics or behaviours can be demographic, sociocultural, economic, etc. It can look at the fertility level, level of education or the unemployment rate of a French-language, English-language or other-mother-tongue population. The characteristics or behaviours of interest can also be linguistic or include a linguistic aspect: bilingualism rate among English-language individuals, average income of workers who use French at work, etc.

To define any population for statistical purposes, inclusion or exclusion criteria must be selected. The choice of criteria defines the border between the target population and the rest of the population, or between the different population groups that make up the total population.^Note The group of interest can therefore be compared with the rest of the population, or different groups can be compared with each other. The groups must be defined before embarking on any comparative analysis. For example, to compare the demographic, language or socioeconomic characteristics of francophones and anglophones, each individual must be placed in their corresponding language group.^Note In this perspective, it is also best to avoid placing individuals in more than one group.

Although the Census of Canada includes several language questions, none of them include an identity dimension or are worded in such a way as to ask about membership in a language group. Consequently, there can be a difference between, on one hand, the preferred definition that reflects the reality we are seeking to describe or represent and, on the other, the possible definitions based on available statistics. These statistics usually limit the possible definitions.^Note Limitations in statistical criteria and classifications can, in turn, have a structuring effect on the groups themselves, such as if they are used to select recipients of particular programs or policies.

The most common variables for defining language groups using census data are generally mother tongue, first official language spoken and language spoken most often at home (or main home language). Other variables, such as knowledge of languages or language of work, can also be used. However, they are more often used as variables of interest to compare such phenomena as the English–French bilingualism rates of different language groups, or the use of languages at work by English-language, French-language or other-mother-tongue workers.

In Canada, there is no consensus on how these language groups, including official language groups, are defined.^Note For many years, the “English” and “French” groups were initially defined by ethnocultural ancestry. Mother tongue was used to measure assimilation, such as by estimating the portion of the population of French ancestry to whom a language other than French had been transmitted as its mother tongue. Following the work of the Royal Commission on Bilingualism and Biculturalism and the adoption of the Official Languages Act by the Government of Canada in 1969, use of mother tongue to indicate membership in an ethnocultural group required a change in the definition of groups based mainly on their language characteristics (Juteau 2015, p. 149). At that time, mother tongue was used to define language groups because it was the only information derived from the Census of Population that made it possible to do so. The only other language information available at the time was knowledge of English or French.

The addition of language questions in the Census of Canada increased opportunities for defining language groups. Considering information on mother tongue (first language learned in childhood and still understood) to be retrospective, the Royal Commission on Bilingualism and Biculturalism suggested adding a question on the “main language of each Canadian” (Royal Commission on Bilingualism and Biculturalism 1967, p. 18). This recommendation led to the addition of a question on language spoken most often at home in the 1971 Census. New questions on knowledge of non-official languages (1991), languages spoken at home on a regular basis in addition to the main language (2001) and languages used at work (two parts) [2001] as well as many possible combinations of language variables resulted in a wide variety of definitions and even in adjustments based on the specific context, issues and needs of data users.

Today, Statistics Canada uses several definitions, usually selected according to the issue at hand. For example, if the analysis is only on official language groups (from an inclusion perspective, i.e., to include as many Canadians as possible in either group), “first official language spoken” (FOLS) is generally used. In contrast, if the analysis is on language diversity or a language group defined according to an Indigenous or immigrant language, the mother tongue criterion is often preferred over the FOLS because there is no category in the latter that identifies these languages.

However, since no language question in the Census of Canada concerns membership or sense of belonging to a language group, criteria that relate to language characteristics, knowledge or practices are used to “objectively” establish or attribute a person’s belonging to a given group. Criteria used to define populations for statistical purposes may not correspond to the identity or sense of belonging of those involved, particularly when the objective characteristics or sense of identity fail to clearly place certain people in a single group. It will always be possible to use subjective criteria to challenge what is used to define or assign “objective” belonging to a group. This is why Statistics Canada does not suggest a standardized definition for language groups or an official definition for “anglophone,” “francophone,” or even “allophone.” The agency recognizes the importance of defining populations or subpopulations according to language criteria and ensures that data users are provided with as much information as possible so they can choose the criteria they deem appropriate. Furthermore, Statistics Canada regularly defines groups or populations based on language criteria in some of the analytical products it disseminates. There are many definitions and they are always contextual.

In the same spirit, Statistics Canada takes no position on how data users define their language groups. The agency can, on request, provide advice and recommendations on what it considers to be the best use of census data, based on the context, but data users decide whether they want to take it into account. Statistics Canada will not invalidate the definition of a language group used by data users unless it is based on faulty data or a design error. In other words, according to Statistics Canada, there is not only one valid way to define language groups. Ultimately, a definition can be appropriate or inappropriate, depending on the context.

2.1.1 Preparing and presenting data: multiple responses

In addition to the criterion underlying their definition, data preparation methods also have an impact on the composition of language groups. A major issue in data presentation is the processing of multiple responses, a growing phenomenon over the past decades. Respondents can know more than one language, speak more than one language at home or use more than one language at work, and even have more than one mother tongue, although the latter is not as common.

Table 5 presents 2016 Census data on mother tongue for the entire Canadian population according to the broad classification.



Table 6 presents the same data according to three simple methods for processing multiple responses. The first method (exclusion) consists of treating multiple responses in a separate category, which basically excludes them from the main groups. This method produces a minimum estimate of the population whose mother tongue is English, French or a language other than English or French. The close to 820,000 respondents who reported more than one mother tongue in 2016 are not included in any of the three main groups.



Conversely, the second method (inclusion) consists of adding all multiple responses that include French to the French mother tongue group and doing the same with the multiple responses that include English and languages other than English or French. In other words, all instances of French, English or languages other than English or French are included in the corresponding language group, regardless of the kind of response, whether it is a single response or a multiple response. The result is a maximum estimate of the population whose mother tongue is English, French or a language other than English or French. Table 7, however, illustrates that the sum of the three populations using the inclusion method exceeds the total Canadian population. The same is true for data on knowledge of languages, or any other variable for which there are too many multiple responses to be redistributed (such as the census variable on ethnic origin). This can be explained by the fact that, when calculating this sum, respondents who report more than one mother tongue are counted two or three times (in the case of three languages – English, French and a Non-official language).



A third method (distribution) produces a mid-point estimate of the population whose mother tongue is English, French or another language. It consists of distributing respondents who reported both English and French as their mother tongue equally between the English and French mother tongue groups, and similarly distributing the responses of type “English and non-official language,” “French and non-official language,” and “English, French and non-official language” (the latter category being distributed equally between the three language groups). The main advantage of this method is the comprehensive distribution of the population into large, mutually exclusive language groups. The distribution of multiple responses between the groups makes it possible to assign a (single) group to respondents for whom the information available does not enable us to say with certainty which group they belong to. Distribution is the only method in which the sum of the language groups equals the total Canadian population without having to use small multiple response groups that serve as residual categories.

The main limitation of the distribution method is that its simplified nature has an artificial aspect that increases with the number of multiple responses. In other words, redistribution works as long as the number of multiple responses remains marginal, since its impact on the data is quite small. Conversely, redistribution would appear to distort the data if the response rates reached a much higher level, such as 50% of the population. There is, however, no predetermined threshold below which the redistribution of multiple responses would be acceptable and above which it would no longer be.^Note

A second limitation of the distribution method is that it enables only a limited number of languages to be considered in data organization and presentation. It thereby rests on an approach that focuses exclusively on Canada’s main official languages.^Note To apply the distribution method to other languages or families of languages in the data tables published on the Statistics Canada website would require exponentially multiplying the response categories in these tables. Table 8 presents the categories needed in a classification to adequately distribute the multiple responses by separating languages other than English or French into two groups; a first containing Indigenous languages and a second containing immigrant languages. This increases the number of multiple response categories from 4 to 10.^Note



Other methods for processing multiple responses can be favoured in certain specific contexts. For example, multiple responses such as “English and non-official language” can be included only in the English mother tongue group and responses like “French and non-official language” only in the the French mother tongue group, whereas responses like "English and French” and “English, French and non-official language” are generally distributed between the English- and French-language groups. There is therefore no more symmetry between the third mother tongue group and the official language groups, and it becomes a residual category that includes only respondents who did not report either of the two official languages. Analytically speaking, this suggests predominance is given to official language groups.

All these methods used to present multiple responses correspond to the traditional data processing and presentation approach that focuses not on the language groups in general, but on the English- and French-language groups in particular. These categorizations were originally used in response to social and political concerns which, historically, related mostly to English- and French-language groups in Canada.^Note The “other,” “non-official” or “allophone” language categories are in fact one very heterogeneous and large residual category that combines Indigenous languages, immigrant languages and sign languages under a single label. This traditional approach has significant limitations with regard to population estimates for a specific mother tongue other than English or French, or for a family of Indigenous or immigrant mother tongues. The classification that contains 10 categories enables to produce population estimates for the English and French mother tongue groups only. The classification that contains 269 categories enables such estimates only according to the exclusion method, since it is impossible to take into account multiple responses (in order to include or distribute them) for specific languages other than English or French.^Note

2.1.2 Languages spoken at home and language of work

These classifications, as well as the multiple response processing methods, also apply to main home language, and can also apply to the language used most often at work. Since 2001, however, both census questions differ from the question on mother tongue due to a second part that asks about languages spoken at home or used at work on a regular basis in addition to the main language.^Note

There are not only more cases where more than one language is spoken at home,^Note but they are also more complex and varied than cases where there is more than one mother tongue. Where more than one language is spoken at home, there can be multiple responses to Part A (when two or three languages are reported to be spoken just as often) or a combination of one or more main languages (Part A) with one or more secondary languages (Part B). For a comprehensive look at languages spoken at home at least on a regular basis, we must cross-tabulate Part A and Part B (see Table 9).



This level of detail and complexity, however, is not necessarily useful when defining a population or subpopulation of interest according to language criteria. In many cases, it is best to take into account only the main language, i.e., the language spoken most often at home (Part A). If we considered both parts, a minimum estimate (exclusion method) similar to the one presented in Table 6 for mother tongue, which took into account respondents who speak only one language at home, would exclude from the French-language population, for example, all persons whose main language is French, but who also speak another language at home, even if it is simply a secondary language. Information on languages regularly spoken at home in addition to the main language, if taken into account in such a way, significantly restricts the population of interest and increases exclusions.^Note

It can also be best to take into account only the main language to produce a mid-point estimate of language groups (distribution method) when the objective is to form mutually exclusive language groups. Using the same example, respondents who speak French only as a secondary language at home are not likely to be included in the French-language group.^Note They will first be included in the group that corresponds to their main language, as long as we are not simply trying to determine whether they speak French at home, but whether they speak it predominantly.

Moreover, a maximum estimate (inclusion method) that considers all languages reported in parts A or B of the question on languages spoken at home provides an overview of language use within the household. In keeping with the same example this method also groups, within a given subpopulation, respondents who speak only French at home and those who speak French only as a secondary language, in combination with one or several other languages. Therefore, when forming language groups based on language spoken at home, there are many more overlaps between subpopulations than for mother tongue, since it is more usual to speak a secondary language at home than to have more than one mother tongue.

With the approach based on mutually exclusive language groups, information on secondary language (Part B) may not seem very useful, and with good reason. However, it paints a more comprehensive and nuanced picture of language practices, particularly for minority languages, be it Indigenous or immigrant languages, or English or French in minority communities. For example, it enables to differentiate complete language transfers (respondents who stop using their mother tongue at home) from partial language transfers (respondents who still speak their mother tongue at home, but only as a secondary language).

Information on secondary language (Part B) can also be useful when defining a population that speaks a given language, regardless of possible overlaps. Doing so, however, eliminates the mutually exclusive nature of language groups which is specific to demolinguistics, by introducing elements that are part of a different approach, one that is based on languages and language practices and characteristics.

2.2 Approach based on languages and language practices and characteristics

The language approach^Note is mostly interested in the language behaviours of the general population, or of a specific subpopulation, without necessarily defining it according to language criteria. This approach meets emerging concerns that stem from certain trends in sociolinguistics or language sociology (Thibault 2001). The aim of these disciplines is much broader than demolinguistics’, with the latter being more closely associated with census data, particularly in Canada^Note (Humbert, Coray and Duchêne 2018). Language approach perspectives can be grouped into, but are not limited to, three main categories.

The first attempts to identify the presence of a language (English, French, an Indigenous language or an immigrant language) within the overall population and measure its evolution. Census data can provide the number of respondents who are able to conduct a conversation in a language, who speak this language at home or use it at work, or who have this language as their mother tongue. The presence of a language within a population can be measured using one of these indicators, or a combination thereof.^Note

The second category could look at language practices and dynamics in a given environment, such as a geographic area (neighbourhood, city, region, etc.), a home environment (family, household, etc.) or a work setting (company, job sector, etc.). Among others, we can examine language diversity in a neighbourhood or region, families or households in which there are several languages, bilingualism or the use of a specific language on the labour market in general or in certain sectors of activity, or even how languages co-exist in the workplace. In this context, the language approach enables us to comprehend language situations in all their complexity.

The third category can look at respondents in a given situation or with specific language characteristics. We can thus study language criteria such as bilingualism and multilingualism, language learning, transmission and loss, and language paths. Moreover, individuals who speak a language, who are bilingual or multilingual, who have undergone a language transfer, who use different languages at home or at work, etc., often captures the attention of researchers and data users. Like the language group approach, the language approach can therefore be used to study specifically a given population or group according to language criteria. In this sense, however, there is no need for the interest groups to be defined in a mutually exclusive manner.

The challenge relating to the processing of multiple responses is different in the case of the language approach. To adequately measure the presence of a language, it is often best to take into account all instances of the language in question. This is particularly true for minority—or very minority—languages, such as certain Indigenous languages. Table 10 indicates that in 2016, multiple response rates were twice as high for non-official languages (8.2%) than for French (3.8%) or English (3.6%) as a mother tongue.



The difference is even greater for statistics on languages spoken at home (see Table 11). About one-quarter of all respondents who speak French at home at least on a regular basis do so in combination with another language, all languages combined. In other words, about three in four respondents who speak French at home do not speak any other language at home. Results are similar among respondents who speak English at home, with a few percentage points difference. These proportions are the opposite among respondents who speak an Indigenous language: more than three-quarters of them (75.9%) speak an Indigenous language in combination with another language.



This is also very true for immigrant languages^Note the integration of non-official-language immigrants in Canadian society is often synonymous with the use of English or French at work, and the gradual adoption of that language as the language used at home.

However, it may not be justified to place all instances of a language on equal footing. Thanks to the two-part questions, census data on languages spoken at home and languages used at work make it possible to differentiate between the single use and the combined use of a language. In the case of the latter, we can also differentiate the (mainly) prevalent use from the equal use and the secondary use of a language.

Yet, when respondents are asked in Part B of these questions to report all languages spoken “regularly” at home or used “regularly” at work in addition to their main language, the threshold according to which a language is spoken or used sufficiently enough to be reported can vary from one person to another. There is no way to certify that respondents will report a language usually spoken or used and not one that is used only occasionally. Nevertheless, qualitative tests performed among respondents and data from the 2006 Survey on the Vitality of Official-Language Minorities (SVOLM) revealed that most respondents consider a language spoken or used “regularly” as one that is used on a daily basis.

Knowledge of a language also comprises an unknown aspect. The degree of language proficiency required to report the ability to conduct a conversation in this language is left at the respondent’s discretion. It is subjective and can vary from one person to another. Furthermore, census data cannot evaluate a person’s language proficiency, nor can it qualitatively differentiate language proficiency among bilingual or multilingual persons. Looking only at the responses to the official language question, we have no way of knowing in which language a bilingual person who speaks English and French is most proficient. Both languages appear to have equal footing in the census data, whereas this is rarely the case in reality.^Note A look at the other language questions can lead to certain hypotheses, but there is no way to be certain.

With regard to mother tongue, the instructions provided in the 2016 Census of Population Long-form Guide specified the following: “For a person who learned two or more languages at the same time in early childhood, report the language this person spoke most often at home before starting school. Report two or more languages only if those languages were used equally often and are still understood by this person.”^Note Multiple responses to the mother tongue question, which are much less frequent than for the other questions, therefore reflect a situation in which languages were transmitted equally to the respondent.

2.2.1 Data preparation and presentation: multiple responses

The different methods for processing multiple responses presented in Section 2.1.1, which are used to form language groups, are not necessarily useful when focusing on a given language. To do so, it is generally best to take into account all respondents who mention that language in response to the selected census question, which corresponds to the maximum estimate (inclusion method) presented in Table 6. For example, when measuring the extent to which the French language is present within a population, with mother tongue as the indicator, it is difficult to justify, based on the pretext that respondents reported this language in combination with another language, the total exclusion (exclusion method) or even partial exclusion (distribution method) of respondents who did in fact report French as their mother tongue, even if multiple responses are somewhat unstable from one census to another (Statistics Canada 2013). In this context, the maximum estimate (inclusion method) is more respectful of the responses given by Canadians to the census question on mother tongue and provides a better overview of the presence of a language in a given society or environment.

To find out, for example, the total number of respondents whose mother tongue is French, we simply need to add up the “French” category from the single responses to the three multiple response categories that include French (see Table 6, inclusion method). For similar information on French spoken at home, the operation is more complex since data are collected using a two-part question. We must first cross-tabulate the information collected in Part A to the information collected in Part B, as in Table 9 (see Section 2.1.2), and add up the information from 20 different cells, that is, all those where French is mentioned for either one of the two parts. In Table 12 below,^Note the grey cells are the ones that must be added up to take into account all respondents who reported speaking French at home at least on a regular basis^Note (without taking into account the order of languages in cases where more than one language is reported).



To simplify this operation while maintaining the distinction between the main (Part A) and secondary (Part B) language behaviours, Statistics Canada used a different approach to present the 2016 Census data on languages spoken at home^Note (see Table 13). The approach consists in presenting, for a given language (or group of languages), the number of respondents who reported using that language (or language belonging to this group) only, mostly, equally or regularly as their secondary language. The variable also indicates the number of respondents who did not report speaking this language (or group of languages) at home at least on a regular basis. It is therefore much easier to have an overview of the use of this language or group of languages at home by Canadians.



The inconvenience with this approach is that data can be presented only for one language or one group of languages at a time. In the 2016 Census, data were presented this way, separately, for French, English, Indigenous languages (combined) and immigrant languages (combined). The same information cannot be obtained for a specific Indigenous language or immigrant language using the data tables currently available on the Statistics Canada website.

2.2.2 Languages other than English and French

Based on the “traditional” data presentation method, multiple response categories are presented in distinct categories for both broad and detailed classification. This provides an overview for only a very limited number of languages, so preference is given to English and French, Canada’s official languages (see Section 2.1.1). It is not possible for data users interested in a specific Indigenous or immigrant language, or a family of languages, to calculate the total number of respondents who reported this language from the tables in which data are presented this way.^Note

To address this issue, some tables^Note published on the Statistics Canada website present mother tongue data the same way as knowledge of languages data (see Section 1.2), taking into account, for each language, all respondents who reported it as their only mother tongue or in combination with one or more languages (multiple responses). Table 14 presents an overview of data that appears on the Statistics Canada website.^Note This is the same information that is presented in Table 10, but with additional detail for each specific language for which census data are available.



The same type of data are also provided for languages spoken at home.^Note In this case, the information is presented for all respondents who reported speaking a given language at least on a regular basis, without being able to determine whether this is their main language (Part A) or secondary language (Part B). When the information is cross-tabulated with the “single and multiple responses for language spoken at home” variable (see Table 15), the “single responses for language spoken at home” category combines all respondents who reported only one language in Part A of the census question on language spoken at home, without having provided a response to Part B. In contrast, the “multiple responses for language spoken at home” combines all respondents who reported more than one language in Part A, or any combination of languages in Part A or B of the question.



2.3 Comparison of both approaches

Language groups are often studied for comparison purposes. This purpose was at the heart of the work by the Royal Commission on Bilingualism and Biculturalism, in which we noticed, among other things, that French-language populations were socioeconomically behind compared with English-language populations.^Note This is the reality in which demolinguistics experienced its largest growth in the 1960 and 1970s, when unequal relationships between language groups were seen by many as a threat to social peace (Martel and Pâquet 2010, p. 20). The comparison can suppose some sort of competition or conflict^Note between the language groups, even though this may not necessarily be the case. The title of the founding paper by Richard J. Joy, Languages in Conflicts: the Canadian Experience, published in 1967, clearly expresses this conflicting aspect in relationships between language groups in Canada.

The language approach does not exclude comparisons between languages or between groups who speak those languages. However, it is best for studying language coexistence, language diversity and multilingualism, among other examples, than the language group approach. But to do so requires a different approach for the presentation and dissemination of language data, particularly for multiple responses.

2.3.1 Mother tongue

Table 16 presents data (counts and proportion of the total population) on mother tongue for Canada, the provinces and territories, according to the distribution and inclusion methods (see Section 2.1.1). The distribution of multiple responses between the major language groups is the method generally used for the language group approach. The inclusion method, which can also be used to define a language group, is the one that best meets the objectives of the language approach.^Note



Of course, the number and proportion of respondents whose mother tongue is English, French or an “other” language are different depending on the approach or method selected. The data are the same; the differences lie only in the handling of the multiple responses. The differences observed on the national, provincial or territorial scale can seem trivial at first: they are less than one percentage point in the majority of cases and barely exceed this threshold in other cases (see Table 16). In numbers, however, these differences translate to 371,890 respondents whose mother tongue is English, 148,335 respondents whose mother tongue is French and 332,305 respondents whose mother tongue is a language other than English or French, across Canada (see Table 17).



In certain provinces, these respondents who are added to or excluded from the population of interest, depending on the point of view, can have quite a significant effect on minority language groups. In Quebec, for example, the redistribution of multiple responses excludes 61,907 respondents who reported English as their mother tongue from the English mother tongue group, which equals 8.6% of all Quebecers who reported English as their mother tongue during the 2016 Census. Similarly, the redistribution of multiple responses excludes 11.2% of respondents who reported French as their mother tongue in Newfoundland and Labrador, and 10.4% of those in British Columbia. All these respondents reported more than one mother tongue.

Based on the language approach, French is the mother tongue of all respondents who respond French to the mother tongue question. Likewise, English is the mother tongue of all respondents who report English as their mother tongue, and so on and so forth for all other mother tongues enumerated. The purpose is not to determine which mother tongue group respondents belong to, but rather how many respondents reported French, English or any other language as their mother tongue.

In the language group approach, the exclusion method and the distribution method are based on the principle that each respondent can belong to only one group. Each respondent therefore has the equivalent of one mother tongue (or 100% of the mother tongue). To a certain extent, respondents who report English and French as their mother tongues have two halves of a mother tongue: 50% of their mother tongue is English and 50% is French, a situation which, in reality, is difficult to justify. They are therefore half counted in the English mother tongue group and half in the French mother tongue group. Statistically speaking, this is the same as placing half of respondents who report English and French as their mother tongue in the French-language group and the other half in the English-language group.

The language approach is dichotomous: respondents reported or did not report a given mother tongue. If so, they are included in the count; if not, they are excluded.^Note As for the language group approach, it attempts to consider the importance of the respondent’s language (or, at least, presupposes lesser or greater importance based on whether there is “coexistence” with another or several other languages). In this sense, we assume that French is of greater importance for respondents who report only this language as their mother tongue, compared with respondents who report it in combination with another language. This importance is measured as follows: 100% for a single response, 50% for a combination of two languages and 33.3% for a combination of three languages.^Note

On the other hand, there is no indication that a person who is transmitted two mother tongues is transmitted only half of each language. The transmission of two mother tongues is done cumulatively, and not to the detriment of one of the languages. With time, a respondent may become more proficient in one language than the other, but census data do not enable us to support this statement. According to the language group approach, reporting only one language has more weight than reporting that language in a multiple response. According to the language approach, all languages reported for a question on mother tongue have the same weight.

2.3.2 Languages spoken at home or languages used at work

The situation is different for languages spoken at home or languages used at work. Table 18 presents data on language spoken at home most often (Part A only) the same way Table 16 presents data on mother tongue, that is, by adding up all respondents who reported English, French or a third language as their main language of use. In this case, respondents report speaking these languages at home in Part A of the question without the responses in Part B being taken into account.



However, Tables 19, 20, 21 and 22 illustrate the importance of taking into account multiple responses and secondary language for a more comprehensive and nuanced picture of the languages spoken at home, particularly in the case of Indigenous languages. Thanks to these tables, users can compare the sum and proportion of respondents who speak, at home, English (Table 19), French (Table 20), an Indigenous language (Table 21) or an immigrant language (Table 22), most often (Part A, single responses), equally (Part A, multiple responses) or regularly (Part B) in addition to their main language. This differentiates prevalent use from the equal or secondary use of a language at home.^Note The total therefore provides an overview of all respondents who speak English, French, an Indigenous language or an immigrant language at home at least on a regular basis.

   

Information on languages spoken at home derived from census data does not make it possible to accurately weigh language use according to all possible combinations, which are more numerous than for mother tongue. Although more weight should be given to main language than to secondary language, nothing indicates the specific weighting that should be attributed to each. For example, if a respondent reports a main language (Part A) and a secondary language (Part B), it is not possible to precisely weigh their respective importance beyond the fact that the main language should have greater weight than the secondary language. What if more than one main language or more than one secondary language is reported? There are too many cases and too much uncertainty to be able to process this information with accuracy. Furthermore, this level of accuracy is not required because the information provided in Part B of the question on languages spoken at home is not generally used to create language groups using a multiple response distribution method (see Section 2.1.2).

However, based on the language approach, data on languages spoken at home and languages used at work provide a nuanced portrait of the use of languages by taking advantage of all the information available. This is why Statistics Canada has developed an approach to distinguish single, prevalent (most often), equal or secondary (regularly) use of a language (see Table 13, Section 2.2.1).

2.4 Data processing

Language data products derived from the Canadian Census of Population have long been prepared and disseminated by Statistics Canada to follow the demographic evolution of language groups, and more specifically English- and French-language groups. It is important for Statistics Canada that data users who want to continue following this evolution are able to do so. The agency will therefore continue to provide data that enable users to define language groups based on criteria they deem relevant and compare the data to previous census data. The 2016 Census data tables published on the Statistics Canada website, like the ones from previous censuses, enable the creation and analysis of mutually exclusive language groups. Data from the 2021 Census will do the same.

However, it is just as important for Statistics Canada to respond to, inasmuch as possible, current and emerging issues that require language statistics to be presented and developed differently. In this sense, the agency would like to develop a data offer that focuses on languages, language practices and language characteristics, as well as on specific languages or families of languages. These data will, of course, concern English and French, but also Indigenous and immigrant languages. This will be done in such a way that does not compromise the tables that data users already had access to in the past.

English and French are Canada’s official languages. Even in a context of growing language diversity, they play a key role in the lives of a vast majority of Canadians. They will continue to play a central role in the presentation of census language data. Offering the data using the language approach improves the statistical information that is available on official languages as much as it does the information available on Indigenous and immigrant languages.

In brief, the objective is not to favour one approach in particular, nor is it to reduce or modify the offer to the detriment of one of the approaches, but rather to develop a complementary data offer so as to meet as many data needs and users as possible.

2.5 Data presentation

Statistics Canada is constantly required to make choices about how it presents data in its analytical documents. The language group approach was used in the past and continues to be used in several documents and analyses. It is still relevant to follow the demographic evolution of language groups in Canada: the relative weight of these groups changes based on steady and very linguistically diverse immigration.^Note Although there are much fewer differences since the findings of the Royal Commission on Bilingualism and Biculturalism in the 1960s,^Note comparing sociocultural and economic characteristics between different language groups meets current concerns related, among others, to specific issues that are often local or regional.^Note It enables to better understand the challenges faced by certain groups. Statistics Canada will continue to disseminate analytical reports that target certain important aspects of language groups, particularly when it comes to official and minority language groups.

The agency also plans to use the approach that focuses on languages and language practices and characteristics in several other documents. In a society that is more and more linguistically diverse, the benefit of the language approach is that it provides a simple, user-friendly and accessible perspective, in addition to respecting responses provided in the Canadian Census of Population, and makes use of all information available. This approach enables to report on growing phenomena within Canadian society, such as bilingualism and multilingualism, both on an individual level and in relation to households, contacts, exchanges and language coexistence.^Note Furthermore, the language approach, based on which all respondents who report a language are taken into account, is better adapted to the study of Indigenous and immigrant languages and the issues that concern them.^Note

Conclusion

Two main approaches can determine how statistical language data derived from the Canadian Census of Population are processed and disseminated. One is based on language groups, and the other on the population’s languages and language practices and characteristics. Without being incompatible, both perspectives are based on different data processing choices as it relates to presenting multiple responses to language questions.

The objective of this paper was to explain both main approaches and their differences. These approaches use different research techniques and therefore provide a different view of the language situation in Canada. The redistribution of multiple responses, usually favoured by the language group approach, has been criticized in the past.^Note In a context where the availability of statistical data on language groups has increased, opportunities for defining language groups have multiplied. The definition of language groups or communities is thus also the responsibility of the groups and communities in question. Statistics Canada’s role is to ensure they have the tools and statistical information required to do so. The language group approach thus continues to be used by numerous researchers and data users because it provides effective responses to certain legitimate questions.

However, this approach does not answer all current questions. The language approach, which considers all respondents who report a language, is better suited to answer certain emerging questions from a language statistics perspective. There is, in fact, growing interest for languages other than English and French, for bilingualism and multilingualism, for more inclusive (and not mutually exclusive) perspectives on languages and populations that know them, speak them or have them as their mother tongue, as well as for a level of detail and nuance that the language group approach cannot easily provide.

For the last few censuses, Statistics Canada has focused more on the approach based on languages, language practices and characteristics of the population in its analytical documents, while continuing to disseminate data that can be used to establish language groups according to the main indicators (mother tongue, main language of use, first official language spoken, etc.). Statistics Canada strives to have an optimal balance between taking into account emerging issues and ensuring historical data comparability. Finding this optimal balance between both sometimes contradictory considerations is at the heart of the agency’s data collection, processing and dissemination activities. Statistics Canada therefore does not favour one approach over another in its data products. Statistics Canada’s objective is to provide the greatest possible number of data users with access to statistical information that is relevant to them, regardless of their chosen approach.

Appendix A

Linguistic questions from the 2016 Census

The following definitions were taken from the 2016 Census Dictionary.

Knowledge of official languages

Definition: 'Knowledge of official languages' refers to whether the person can conduct a conversation in English only, French only, in both or in neither language. For a child who has not yet learned to speak, this includes languages that the child is learning to speak at home.

Question of the census:

Description for figure A.1

7 Can this person speak English or French well enough to conduct a conversation?

Mark " X " one circle only.

Language spoken at home

Language spoken most often at home

Definition: 'Language spoken most often at home' refers to the language the person speaks most often at home at the time of data collection. A person can report more than one language as 'spoken most often at home' if the languages are spoken equally often.

For a person who lives alone, the language spoken most often at home is the language in which he or she feels most comfortable. For a child who has not yet learned to speak, this is the language spoken most often to the child at home. Where two languages are spoken to the child, the language spoken most often at home is the language spoken most often. If both languages are used equally often, then both languages are included here.

Other language(s) spoken regularly at home

Definition: 'Other language(s) spoken regularly at home' refers to the languages, if any, that the person speaks at home on a regular basis at the time of data collection, other than the language or languages he or she speaks most often at home.

Questions from the census:

Description for figure A.2

8 a) What language does this person speak most often at home?

b) Does this person speak any other languages on a regular basis at home?

Mother tongue

Definition: 'Mother tongue' refers to the first language learned at home in childhood and still understood by the person at the time the data was collected. If the person no longer understands the first language learned, the mother tongue is the second language learned. For a person who learned two languages at the same time in early childhood, the mother tongue is the language this person spoke most often at home before starting school. The person has two mother tongues only if the two languages were used equally often and are still understood by the person. For a child who has not yet learned to speak, the mother tongue is the language spoken most often to this child at home. The child has two mother tongues only if both languages are spoken equally often so that the child learns both languages at the same time.

Description for figure A.3

9 What is the language that this person first learned at home in childhood and still understands?

If this person no longer understands the first language learned, indicate the second language learned.

Knowledge of non-official languages

Definition: 'Knowledge of non-official languages' refers to whether the person can conduct a conversation in a language other than English or French. For a child who has not yet learned to speak, this includes languages that the child is learning to speak at home. The number of languages that can be reported may vary between surveys, depending on the objectives of the survey.

Question from the census:

Description for figure A.4

16 What language(s), other than English or French, can this person speak well enough to conduct a conversation?

Language at work

Language used most often at work

Definition: 'Language used most often at work' refers to the language the person uses most often at work. A person can report more than one language as "used most often at work" if the languages are used equally often

Other language(s) used regularly at work

Definition: 'Other language(s) used regularly at work' refers to the languages, if any, that the person uses in their job on a regular basis, other than the language or languages he or she uses most often at work.

Question from the census:

Description for figure A.5

45 a) In this job, what language did this person use most often?

b) Did this person use any other languages on a regular basis in this job?

Appendix B

Detailed classification

• Total - Mother tongue
  • Single responses
    • Official languages
      • English
      • French
    • Non-official languages
      • Aboriginal languages
        • Algonquian languages
        • Blackfoot
        • Cree-Montagnais languages
          • Atikamekw
          • Montagnais (Innu)
          • Moose Cree
          • Naskapi
          • Northern East Cree
          • Plains Cree/li>
          • Southern East Cree
          • Swampy Cree
          • Woods Cree
          • Cree, n.o.s.
        • Eastern Algonquian languages
          • Malecite
          • Mi'kmaq
        • Ojibway-Potawatomi languages
          • Algonquin
          • Ojibway
          • Oji-Cree
          • Ottawa (Odawa)
        • Algonquian languages, n.i.e.
      • Athabaskan languages
        • Northern Athabaskan languages
          • Babine (Wetsuwet'en)
          • Beaver
          • Carrier
          • Chilcotin
          • Dene
          • Dogrib (Tlicho)
          • Gwich'in
          • Sarsi (Sarcee)
          • Sekani
          • Slavey-Hare languages
            • North Slavey (Hare)
            • South Slavey
            • Slavey, n.o.s.
          • Tahltan languages
            • Kaska (Nahani)
            • Tahltan
          • Tutchone languages
            • Northern Tutchone
            • Southern Tutchone
        • Athabaskan languages, n.i.e.
      • Haida
      • Inuit languages
        • Inuinnaqtun (Inuvialuktun)
        • Inuktitut
        • Inuit languages, n.i.e.
      • Iroquoian languages
        • Cayuga
        • Mohawk
        • Oneida
        • Iroquoian languages, n.i.e.
      • Kutenai
      • Michif
      • Salish languages
        • Comox
        • Halkomelem
        • Lillooet
        • Okanagan
        • Shuswap (Secwepemctsin)
        • Squamish
        • Straits
        • Thompson (Ntlakapamux)
        • Salish languages, n.i.e.
      • Siouan languages
        • Dakota
        • Stoney
        • Siouan languages, n.i.e.
      • Tlingit
      • Tsimshian languages
        • Gitxsan (Gitksan)
        • Nisga'a
        • Tsimshian
      • Wakashan languages
        • Haisla
        • Heiltsuk
        • Kwakiutl (Kwak'wala)
        • Nuu-chah-nulth (Nootka)
        • Wakashan languages, n.i.e.
      • Aboriginal languages, n.o.s.
    • Non-Aboriginal languages
      • Afro-Asiatic languages
        • Berber languages
          • Kabyle
          • Berber languages, n.i.e.
        • Cushitic languages
          • Bilen
          • Oromo
          • Somali
          • Cushitic languages, n.i.e.
        • Semitic languages
          • Amharic
          • Arabic
          • Assyrian Neo-Aramaic
          • Chaldean Neo-Aramaic
          • Harari
          • Hebrew
          • Maltese
          • Tigrigna
          • Semitic languages, n.i.e.
        • Afro-Asiatic languages, n.i.e.
      • Austro-Asiatic languages
        • Khmer (Cambodian)
        • Vietnamese
        • Austro-Asiatic languages, n.i.e
      • Austronesian languages
        • Bikol
        • Cebuano
        • Fijian
        • Hiligaynon
        • Ilocano
        • Malagasy
        • Malay
        • Pampangan (Kapampangan, Pampango)
        • Pangasinan
        • Tagalog (Pilipino, Filipino)
        • Waray-Waray
        • Austronesian languages, n.i.e.
      • Creole languages
        • Haitian Creole
        • Creole, n.o.s.
        • Creole languages, n.i.e.
      • Dravidian languages
        • Kannada
        • Malayalam
        • Tamil
        • Telugu
        • Dravidian languages, n.i.e.
      • Hmong-Mien languages
      • Indo-European languages
        • Albanian
        • Armenian
        • Balto-Slavic languages
          • Baltic languages
            • Latvian
            • Lithuanian
          • Slavic languages
            • Belarusan
            • Bosnian
            • Bulgarian
            • Croatian
            • Czech
            • Macedonian
            • Polish
            • Russian
            • Serbian
            • Serbo-Croatian
            • Slovak
            • Slovene (Slovenian)
            • Ukrainian
            • Slavic languages, n.i.e.
          • Celtic languages
            • Scottish Gaelic
            • Welsh
            • Celtic languages, n.i.e.
          • Germanic languages
            • Afrikaans
            • Danish
            • Dutch
            • Frisian
            • German
            • Icelandic
            • Norwegian
            • Swedish
            • Vlaams (Flemish)
            • Yiddish
            • Germanic languages, n.i.e.
          • Greek
          • Indo-Iranian languages
            • Indo-Aryan languages
              • Bengali
              • Gujarati
              • Hindi
              • Kashmiri
              • Konkani
              • Marathi
              • Nepali
              • Oriya (Odia)
              • Punjabi (Panjabi)
              • Sindhi
              • Sinhala (Sinhalese)
              • Urdu
            • Iranian languages
              • Kurdish
              • Pashto
              • Persian (Farsi)
            • Indo-Iranian languages, n.i.e.
          • Italic (Romance) languages
            • Catalan
            • Italian
            • Portuguese
            • Romanian
            • Spanish
            • Italic (Romance) languages, n.i.e.
          • Japanese
          • Kartvelian languages
            • Georgian
          • Korean
          • Mongolic languages
            • Mongolian
          • Niger-Congo languages
            • Akan (Twi)
            • Bamanankan
            • Edo
            • Ewe
            • Fulah (Pular, Pulaar, Fulfulde)
            • Ga
            • Ganda
            • Igbo
            • Lingala
            • Rundi (Kirundi)
            • Kinyarwanda (Rwanda)
            • Shona
            • Swahili
            • Wolof
            • Yoruba
            • Niger-Congo languages, n.i.e.
          • Nilo-Saharan languages
            • Dinka
            • Nilo-Saharan languages, n.i.e.
          • Sign languages
            • American Sign Language
            • Quebec Sign Language
            • Sign languages, n.i.e
          • Sino-Tibetan languages
            • Chinese languages
              • Cantonese
              • Hakka
              • Mandarin
              • Min Dong
              • Min Nan (Chaochow, Teochow, Fukien, Taiwanese)
              • Wu (Shanghainese)
              • Chinese, n.o.s.
              • Chinese languages, n.i.e.
            • Tibeto-Burman languages
              • Burmese
              • Karenic languages
              • Tibetan
              • Tibeto-Burman languages, n.i.e.
            • Tai-Kadai languages
              • Lao
              • Thai
              • Tai-Kadai languages, n.i.e
            • Turkic languages
              • Azerbaijani
              • Turkish
              • Uyghur
              • Uzbek
              • Turkic languages, n.i.e.
            • Uralic languages
              • Estonian
              • Finnish
              • Hungarian
              • Uralic languages, n.i.e.
            • Other languages, n.i.e.
          • Multiple responses
            • English and French
            • English and non-official language
            • French and non-official language

References

Arsenault Morin, Alex and Vincent Geloso. 2019. "Multilingualism and the decline of French in Quebec". Journal of multilingual and multicultural development. October 22. https://www.tandfonline.com/doi/abs/10.1080/01434632.2019.1676250

Béland, Paul. 2008. Langue et immigration, langue du travail : éléments d’analyse. Conseil supérieur de la langue française. 40 p. http://www.cslf.gouv.qc.ca/publications/pubf228/f228.pdf

Bérard-Chagnon, Julien and Jean-François Lepage. 2016. "The literacy skills of New Brunswick francophones: Demographic and socioeconomic issues" Ethnicity, Language and Immigration Thematic Series. Statistics Canada Catalogue no. 89- 657-X2016001. 80 p. https://www150.statcan.gc.ca/n1/en/pub/89-657-x/89-657-x2016001-eng.pdf?st=9tF8ENG8

Boyer, Henri. 2017. Introduction à la sociolinguistique. Dunod. 136 p.

Corbeil, Jean-Pierre. Forthcoming. “Catégories et frontières : le recensement et la construction sociale, politique et scientifique des groupes ethnolinguistiques au Canada.” Diversité urbaine.

Houle, René and Jean-Pierre Corbeil. 2017. Language Projections for Canada, 2011 to 2036. Ottawa. Statistics Canada Catalogue no. 89-657-X2017001. 140 p. https://www150.statcan.gc.ca/n1/en/catalogue/89-657-X2017001

Humbert, Philippe, Renata Coray and Alexandre Duchêne. 2018. Compter les langues : histoire, méthodes et politiques des recensements de population. Une revue de littérature. Fribourg. Research Centre on Multilingualism. 73 p. http://doc.rero.ch/record/308744/files/2018.001_census_def_web_0.pdf

Joy, Richard J. 1967. Languages in Conflict: The Canadian Experience. Ottawa. Self published. 145 p.

Juteau, Danielle. 2015. L’ethnicité et ses frontières. Les Presses de l’Université de Montréal. Montréal. 306 p.

Lachapelle, Réjean. 1991. Utilisation des données de recensement dans la mise en œuvre de la législation linguistique. Ottawa. Statistics Canada. 54 p.

Lachapelle, Réjean and Jean-François Lepage. 2010. Languages in Canada: 2006 Census. Canadian Heritage and Statistics Canada. 205 p.

Lepage, Jean-François, Stéphanie Langlois and Martin Turcotte. 2019. Evolution of the language situation in Nunavut, 2001 to 2016. Ottawa. Statistics Canada Catalogue no. 89-657-X2019010. 102 p. https://www150.statcan.gc.ca/n1/pub/89-657-x/89-657-x2019010-eng.htm

Maheu, Robert. 1985. “La démolinguistique : objet et enjeux.” Cahiers québécois de démographie. Vol. 14, no. 1. April 1985. p. 3-5. https://www.erudit.org/fr/revues/cqd/1985-v14-n1-cqd2458/600554ar/

Martel, Marcel and Martin Pâquet. 2010. Langue et politique au Canada et au Québec. Une synthèse historique. Boréal. Montréal. 335 p.

Royal Commission on Bilingualism and Biculturalism. 1967. “Book I, Official Languages.” Report of the Royal Commission on Bilingualism and Biculturalism. Ottawa. 229 p. http://publications.gc.ca/site/eng/9.643815/publication.html

Royal Commission on Bilingualism and Biculturalism. 1969. “Book III, The Work World.” Report of the Royal Commission on Bilingualism and Biculturalism. Ottawa. 492 p. http://publications.gc.ca/site/eng/9.643817/publication.html

Statistics Canada. 1992. Population selon la langue maternelle, 1981 à 1991 : données de base et statistiques comparatives approchées. Draft document. Ottawa. 15 p.

Statistics Canada. 2013. Methodological Document on the 2011 Census. Ottawa. Statistics Canada Catalogue no. 98-314-X2011051. 30 p.

Statistics Canada. 2017. Linguistic diversity and multilingualism in Canadian homes. Ottawa. Statistics Canada Catalogue no. 98-200-X2016010. 9 p. https://www12.statcan.gc.ca/census-recensement/2016/as-sa/98-200-x/2016010/98-200-x2016010-eng.cfm

Thibault, Pierrette. 2001. “Regard rétrospectif sur la sociolinguistique québécoise et canadienne.” Revue québécoise de linguistique. Vol. 30, no 1. p 19-42. https://www.erudit.org/fr/revues/rql/2001-v30-n1-rql411/000512ar/