The hidden face of public language policy: a case study from the UK

ABSTRACT

In which domains and for which language types does language legislation occur and how easy is it to identify it? The United Kingdom (UK) affords a good test bed to answer these questions since it is often considered to be lacking in strong public language policy due to, amongst other things, a lack of a coherent language policy across the UK and its devolved administrations and a perceived societal disinterest in languages. Through analysis of a corpus of primary and secondary legislation from the UK and its constituent jurisdictions that contain stipulations about language(s), this article shows that UK language legislation spans multiple domains, including public health and safety, law and crime, transport and the media. Whilst some of the legislation, such as the Welsh Language Acts, explicitly deals with language(s), the vast majority of the UK's legislation which contains provisions concerning language(s) is hidden in legislation which primarily concerns another domain. Although hidden, at times these language stipulations mark important landmarks in the status of languages. All this has consequences for the UK language policy landscape, potentially diminishing the perceived importance of languages in and to government and affecting policymakers' ability to collaborate across government in a coherent way.

KEYWORDS:

• Language legislation
• United Kingdom
• explicit vs hidden
• language type
• domain

Introduction

In which domains and for which language types does language legislation occur and how easy is it to identify it? Regarding the first question, we might expect, for instance, much of the legislation which concerns languages to cluster in the education domain, but are there also provisions for languages in domains we might not immediately associate with language issues? In relation to our second question, does language legislation primarily concern the minoritized indigenous languages, community or heritage languages, or the acquisition of other languages? We will argue in this article that the answers to these two questions are coloured by how we identify relevant legislation and that if we are more flexible in our interpretation of ‘language legislation’ to include all legislation which includes clauses concerning languages, regardless of the primary focus of that legislation—what we will term hidden legislation—then different conclusions emerge.

The United Kingdom (UK) provides an interesting test bed for these questions since it is often considered not to have a strong public language policy or indeed, as a predominantly English-speaking nation, much interest in languages—be they indigenous languages (including British Sign Language (BSL), Cornish, Irish, Manx, Scottish Gaelic, Scots, Ulster Scots, Welsh), community or heritage languages, or those acquired through formal language learning. A 2019 call for action from the four UK national academies declared that ‘The UK is currently nowhere near to fulfilling its linguistic potential’ and pressed for a national strategy to remedy this (British Academy et al., 2019, pp. 2–4). We focus here exclusively on primary and secondary legislation from the UK and the devolved administrations, which proved much more numerous than anticipated; the analysis of non-legislative UK language policy documents offers a fruitful avenue for future research.

There are, naturally, some very obvious examples of explicit language policies, notably in Scotland and Wales, such as the Welsh Language Act 1993,¹ and the Gaelic Language (Scotland) Act 2005,² which legislate for the promotion and facilitation of use of Welsh and Gaelic. The situation in Northern Ireland is more complicated because of the association of Irish with the Catholic, Nationalist and Republican community (see below). Nevertheless, there is a general perception of a relative lack of UK public language policy, especially when compared with elsewhere in Europe; for instance, in France the discussion of politique linguistique has a much higher profile, both in government and wider public discourse (Ayres-Bennett, forthcoming). There are a number of reasons for this perception. First, whilst there have been recent initiatives to improve cross-Whitehall collaboration on languages, the UK is still far from having a coherent joined-up language policy, as was already highlighted by the Nuffield Languages Inquiry (2000, p. 6) over twenty years ago. This has led Ayres-Bennett and others to argue that ‘languages are everywhere and nowhere in government’ (Ayres-Bennett, 2017) and, as we will show, this is also true of legislation relating to language(s) in the UK.

A second reason is that it is often assumed that languages in the UK are predominantly a matter for the Department for Education (DfE) and that legislation relating to languages primarily concerns modern languages (ML) education. This conjecture may derive in part—especially in England—from the perceived monolingualism in English of the UK and the accompanying assumption that language policy must therefore relate to the teaching of other languages in schools. As we will see, there is, indeed, explicit legislation for languages in this area, e.g. The Education (National Curriculum) (Attainment Targets and Programmes of Study in Modern Foreign Languages) Order 1995,³ which sets attainment targets and programmes of study in ML for Key Stage (KS)3 and KS4 pupils (aged 11–16) in England and Wales. However, the scope of public language policy, as we will show, is far broader even within the DfE.

A third factor relates to the way language legislation is often hedged with caveats and therefore open to interpretation and non-implementation. For instance, whilst there is legislative provision for the translation of election information for voters and school policy documents for parents, this may be expressed using modal verbs expressing possibility such as ‘may’ rather than obligation, or with hedging clauses, e.g. information will be translated ‘as appears to the governors or authority to be appropriate’.⁴

A final, highly significant reason will be our main focus. We will show that, contrary to general perceptions, legislation relating to languages emanates from a wide range of Ministries and Departments in Whitehall and the devolved administrations, and that it covers community languages, indigenous languages and ML educational policy. Nevertheless, in being frequently embedded in public policy relating primarily to another domain (e.g. Transport or Finance), much of it remains hidden, not just from the public, but from the policymakers themselves.

The article will be structured as follows: first, we introduce the UK legislative context; next, we explain the methodology used to create our corpus of UK language legislation. We then present the findings of our analysis according to the following parameters: explicit vs hidden legislation; legislation type (primary or secondary); jurisdiction; language type (community, indigenous and ML); and domain, before concluding.

Legislative context

Since 1998, certain powers in Northern Ireland (NI), Scotland and Wales have been devolved from the UK Government and Parliament, creating distinct governments and parliaments or assemblies in each jurisdiction: the Northern Ireland Assembly and Northern Ireland Executive; the Scottish Parliament and Scottish Government; and Senedd Cymru and Welsh Government. In simple terms, the Northern Ireland Act 1998,⁵ the Scotland Act 1998,⁶ and the Government of Wales Act 1998,⁷ respectively set out the devolution settlement for each area and the remit and powers of the devolved administration, each with its own devolved and reserved powers (Table 1). The UK Parliament is able to pass primary (UK Public General Acts) and secondary legislation (UK Statutory Instruments) in the reserved policy areas which apply to either the UK as a whole or to one or more jurisdictions. In Scotland, the Scottish Parliament has been able to pass primary legislation (Acts of the Scottish Parliament) and secondary legislation (Scottish Statutory Instruments) in devolved areas since devolution. Further powers were assented in the Scotland Act 2012,⁸ and Scotland Act 2016,⁹ including greater control over the administration of elections and taxation.

Table 1. Devolved and reserved policy areas.

Policy Area	Scotland	Wales	Northern Ireland
Health and social care	D	D	D
Education and training	D	D	D
Local Government	D	D	D
Agriculture, forestry and fisheries	D	D	D
Transport	D	D	D
Some taxation	D	D	D
Justice and Policing	D	R	D
Some social security elements	D	R	D
Sports and the arts	D	D	D
Defence	R
Foreign Affairs
Immigration
Trade Policy
Constitution
Broadcasting (NI may legislate with SoS consent)

Source: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/770709/DevolutionFactsheet.pdf

In Wales, the Government of Wales Act 1998 provided a legal basis for the National Assembly for Wales and allowed it to make secondary legislation.¹⁰ The power to make primary legislation in devolved policy areas, in the form of Measures of the National Assembly for Wales, was granted by the Government of Wales Act 2006,¹¹ which also established the Welsh Assembly. However, Measures could not be passed without approval from other legislative authorities including the UK Parliament. Following a 2011 referendum, the National Assembly was given the power to pass legislation in more policy areas and primary legislation (Acts of the National Assembly for Wales). In 2020, the National Assembly became Senedd Cymru (Welsh Parliament), reflecting the Parliament’s enhanced legislative powers, with primary legislation known as Acts of Senedd Cymru.¹²

The situation for NI is more complicated. From 1921, until its devolved powers were suspended in 1972, NI was a distinct jurisdiction, able to make its own primary legislation (Acts of the Parliament of Northern Ireland). During direct rule, which continued until the Northern Ireland Act 1998, two consecutive Northern Ireland assemblies were established, the first in 1974, able to enact primary legislation in the form of Measures of the Northern Ireland Assembly, the second in 1982 without legislative powers (House of Commons Information Office, 2010, pp. 1–2). Following the 1998 Act, certain legislative powers were transferred from Whitehall to the Northern Ireland Assembly, although it was suspended October 2002-May 2007 and January 2017-January 2020. Since May 2022, it has once again been unable to function since, following elections, it failed to elect a speaker.

Educational institutions are sites in which language policy is particularly evident and, indeed, powerful (Spolsky, 2009, p. 90). While Education and training are now devolved in all three jurisdictions, educational policy in England, NI, Scotland, and Wales developed separately to differing degrees well before devolution.¹³ Scotland’s educational system has been managed independently since 1707 (Clark, 1997, p. 3). In 1921, a Ministry of Education for NI was created and took control of all public schooling except universities.¹⁴ The Welsh education system was traditionally closely aligned to the English system until, with the Education Act of 1902 and the creation of a Welsh Department in the Board of Education in 1907, it slowly began to diverge. Following devolution, the Welsh Government took almost full control of the education system (Jones, 2006, p. 273).

Since devolution, ML policy has increasingly diverged across the four constituent nations (see Ayres-Bennett & Carruthers, 2019). In England, languages are a statutory requirement of the National Curriculum at KS2 and KS3 (ages 7–14); the requirement to continue language study in KS4 in England, as well as in NI and Wales, was removed in 2004.¹⁵ In 2012, the Scottish Government set out to implement a 1+2 policy for language education (based on the model recommended by the European Commission 2004) with the aim of full implementation by August 2021 (Scottish Government, 2012), delayed a year because of Covid-19. This policy gives every child in Scotland the entitlement to learn—in addition to a first language (L1), usually English or BSL—an additional language (L2) from the first year of primary education (P1) to the end of S3 (pupils aged 5–14), and an L3 during the last three years of primary school and at least one year of senior school (in practice usually S2 or S3).

In Wales, educational language policy has prioritized increasing the number of Welsh speakers, through improving teaching and learning of Welsh in schools and developing and increasing Welsh-medium teaching, with a target of one million Welsh speakers by 2050 (Welsh Government, 2017). ML education in Wales has fared less well, with the number of pupils taking public examinations in languages in decline (Gorrara et al., 2020, p. 245).¹⁶ Similarly to Scotland, the Welsh Government’s (2015) Global futures strategic plan aimed to teach pupils three languages, but using a ‘bilingual plus 1’ approach (English, Welsh and one other ML). A new curriculum,¹⁷ introduced in 2022, lists ‘Languages, Literacy and Communication’ as one of six key areas of learning. The area brings together the teaching of English and Welsh from the age of 3, and an ‘international language’ beginning in primary school.

Unlike the rest of the UK, in NI there is currently no statutory provision for teaching languages at primary level, and actual provision is patchy (Carruthers & Ó Mainnín, 2018, p. 161). In secondary education, language education is compulsory at KS3 only, the shortest period in Europe (Collen, 2021, p. 5). Provision for Irish-medium education is growing; the promotion of education in the Irish language is legislated for in the Education (Northern Ireland) Order 1998 (Carruthers & Ó Mainnín, 2018, p. 161). Language policy in NI is politically sensitive. To give some very brief historical context, the establishment of NI in 1922 ignited tensions between the Protestant and Catholic communities in the state and resulted in a 30-year conflict from 1968—known as the Troubles (Dunlevy, 2020, p. 2). The Irish language is intrinsically linked to other societal divisions and the question of an Irish language Act has been divisive, being a contributing factor to the collapse of the Stormont Assembly in 2017 (Dunlevy, 2020, pp. 1–2).

The teaching of community languages often forms little or no part of curricula in mainstream UK schools, their teaching falling instead to the strong network of volunteer-run complementary or ‘Saturday’ schools (Creese & Martin, 2006, p. 1; Hancock & Hancock, 2019, p. 3). These schools, ‘set up in response to the failure of the mainstream education system to meet the needs of ethnic minority children and their communities’ (Wei, 2006, p. 78), not only give pupils the opportunity to learn and develop a community language, but also to participate in cultural activities with those from shared backgrounds (Hancock & Hancock, 2019, pp. 10–11). We shall see that community-language speakers are offered some linguistic protections in UK legislation, but, in our corpus, there is no legislation relating to community languages education.

As already indicated, there are two types of UK legislation—primary and secondary. Primary legislation describes ‘the main laws passed by the legislative bodies of the UK’.¹⁸ Secondary legislation is ‘delegated legislation made by a person or body under authority contained in primary legislation’.¹⁹ There are numerous forms of secondary legislation but those relevant here are Statutory Instruments from the UK and devolved administrations (termed Statutory Rules in NI).

Methodology

Our corpus comprises primary and secondary legislation, January 1918-June 2021, and can be consulted in the repository of language policy documents created as part of the Promoting Language Policy project funded by the Arts and Humanities Research Council.²⁰ Legislation concerning the UK’s indigenous languages, community languages and ML was collected from the government’s online legislation database.²¹ Two forms of legislation were excluded: draft legislation and bills (we focus on legislation which is, or has been, in force) and commencement orders (which bring into force legislation or parts of legislation).

We chose 1918 as the terminus a quo, since it witnessed the publication of the 1918 Education Act²² and the landmark Leathes report (Leathes 1928 [1918]). The former raised the school leaving age to 14 and abolished elementary school fees, marking the beginning of ML education in England and Wales. The latter government-commissioned report examined the position of ML in the British educational system.

Legislation concerning languages was located using the database’s Search function, which allows key term searches of both legislation titles and content. This gave access to legislation which explicitly concerns language, e.g. with ‘language’ or ‘translation’ in the title, and to legislation primarily on another topic, e.g. the Radioactive Material (Road Transport) (Great Britain) Regulations 1996 (discussed below),²³ but which contains one or more articles which make rulings or statements about language(s). The search terms were as follows: ‘Cornish’; ‘Gaelic’; ‘in English’; ‘in Welsh’; ‘Irish’; ‘language’; ‘lingu*’; ‘Manx’; ‘multiling*’; ‘sign language’; ‘Standard English’; ‘Standard Welsh’; ‘Ulster Scots’. The terms ‘English’ and ‘Welsh’ were not used since these generated very many irrelevant contexts: for instance, all Welsh Statutory Instruments contain the word ‘Welsh’. The problem did not arise, however, for other indigenous languages, e.g. Gaelic, which when used as a search term returned fewer results and a larger majority of legislation concerning the Gaelic language. Once located, the legislation was checked for relevance. Legislation which did not concern natural languages was removed, e.g. The Export of Goods (Control) Order 1992²⁴ contains the word ‘language’ three times but in each case this refers to a computer language.

Once collected, legislation was categorized, first, by jurisdiction (UK/England/NI/Scotland/Wales) and broad type (primary/secondary). Both pre- and post-devolution legislation is included. Pre-devolution legislation applying to the UK as a whole is included in the UK list, e.g. The Toys (Safety) Regulations 1989.²⁵ However, pre-devolution legislation whose remit is limited to certain areas is listed alongside post-devolution legislation for that area, e.g. The Non-Domestic Rating (Demand Notices) (Wales) Regulations 1993 (UK SI).²⁶ Similarly, post-devolution legislation limited to individual administrations is included in that area’s list, rather than the UK list, e.g. The National Assembly for Wales (Legislative Competence) (Welsh Language) Order 2010 (UK SI).²⁷

The legislation was further categorized by language type (indigenous/community/ML).²⁸ Legislation can relate to more than one language type, e.g. The Education (Bursaries for Teacher Training) (Amendment) Regulations 1990²⁹ amend the financial provisions for the training of ML and Welsh teachers and therefore concern both ML and indigenous languages. Consequently, percentages are not given for language type, just raw figures. Not all language legislation concerns one of these language types, e.g. legislation concerning English as an additional language (EAL), discussed below in hidden secondary legislation, implicitly refers to speakers of community languages but concerns the provision of English-language teaching.

A classification based on domain was then applied (Table 2). The development of the classification was primarily data-driven but efforts were made to base the categories broadly on domains mapping onto public policy areas/government departments. This allowed quantitative and qualitative analysis of the principal domains in which language is legislated for, both explicitly and implicitly. Decisions around the placement of legislation in categories are not always clear-cut; some legislation arguably fits into more than one category. Data-driven classifications, although not unflawed, are intended to combat the bias of researchers forcing data into pre-determined categories, instead allowing the data itself to drive these. Whilst this bias cannot be ruled out completely, both researchers were actively involved in categorizing the legislation, and before embarking on the main analysis a series of independent categorization exercises with cross-checking were conducted to ensure as far as possible consistency and rigour. Once categorized, lists for each devolved administration were sent to relevant policymakers to be checked for comprehensiveness.

Table 2. Legislation domains.

Armed forces

Business

Census

Charities

Education

Elections and voting

EU law and Brexit

Finance

Government

Health services

Land and property

Law and crime

Media

Nationality and citizenship

Public health and safety

Registration of births, deaths and marriages

Social care

Status, protection and promotion of languages

Tax

Transport

Other

A final important distinction made is between what we term explicit language legislation and more hidden legislation. We use the term ‘explicit’ language legislation to refer to legislation whose primary focus is a language issue, typically concerning the status, protection or promotion of languages. The term ‘hidden’ legislation is used to refer to legislation the primary focus of which is another domain, but which nevertheless includes one or more clauses relating to language issues. For instance, the Scottish Land Court Act 1993 stipulates that ‘one of the members of the Land Court shall be a person who can speak the Gaelic language’ (Art.5).³⁰ This hidden legislation equally contributes to the UK’s language policy landscape and in some cases, as we shall see, represents important landmarks in the status of languages. Since we are interested particularly in demonstrating that much language legislation is hidden, we use this distinction to structure our quantitative and qualitative analysis. It is important to note, in the quantitative analysis, that the frequency figures given relate only to the frequency within our corpus; they do not take account of the proportion of all legislation devoted to each domain.

To summarize, the following classifications were applied:

1. Explicit vs hidden

2. Jurisdiction (UK/England/NI/Scotland/Wales)³¹

3. Legislation type (primary/secondary)

4. Language type (indigenous/community/ML)

5. Domain

Results and discussion

Overall trends

Our corpus comprises 1,501 pieces of legislation. Some themes considered only marginal to language policy are not included in the analysis but can be found in separate lists in the online language policy repository: communication and comprehension difficulties; the removal of offensive language or the injunction to use ordinary language; agreements which are either international, apply to overseas territories or concern relations between specific nations; the requirement to speak English for certain professions and functions; and the listing of specific nomenclature, typically in one of the indigenous languages. Legislation outlining devolution settlements was also excluded.

Table 3 gives the categorization of all legislation in the corpus.³² Looking at the full corpus breakdown, it is unsurprising that the UK is generally believed to have little or no language policy; 92.8% (n = 1,393) of legislation which references language(s) is hidden in legislation predominantly related to other domains. We also clearly see that whilst there is legislation concerning ML, there is far more relating to indigenous and community languages, dispelling another common belief that UK language policy primarily concerns ML education. Furthermore, whilst there is a substantial amount of linguistic legislation relating to Education (n = 283, 18.9%), 81.1% of legislation relates to other domains, and more legislation relates to Public health and safety (n = 316, 21.1%). In fact, language legislation spans 21 domains, confirming that it concerns much more than simply language status and education. In the following sections we explore our categorizations in more detail.

Table 3. Full corpus categorization.

		Number	% of corpus
Explicit vs hidden	Explicit	108	7.2%
	Hidden	1,393	92.8%
Jurisdiction	UK	583	38.8%
	Wales	240	16.0%
	Scotland	215	14.3%
	NI	179	11.9%
	England	152	10.1%
	England and Wales	128	8.5%
	England, Wales and NI	2	0.1%
	England, Wales and Scotland	2	0.1%
Legislation type	Primary	193	12.9%
	Secondary	1,308	87.1%
Language type	Indigenous	686	N/A
	Community	270	N/A
	ML	90	N/A
Domain	Public health and safety	316	21.1%
	Education	283	18.9%
	Law and crime	197	13.1%
	EU law and Brexit	114	7.6%
	Social care	92	6.1%
	Business	84	5.6%
	Finance	82	5.5%
	Health services	53	3.5%
	Elections and voting	49	3.3%
	Transport	47	3.1%
	Registration of births, deaths and marriages	28	1.9%
	Status, protection and promotion of languages	28	1.9%
	Media	26	1.7%
	Government	23	1.5%
	Nationality and citizenship	20	1.3%
	Land and property	19	1.3%
	Census	15	1.0%
	Tax	4	0.3%
	Charities	3	0.2%
	Armed forces	2	0.1%
	Other^a	16	1.1%

^a The Other domain—all of which has a hidden language element—contains legislation which did not fit readily into any other domain. Examples include Intelligences Services Act 1994; Disability Discrimination Act 1995; The Pet Travel Scheme (Pilot Arrangements) (England) Order 1999; The Cremation (England and Wales) (Amendment) Regulations 2017.

Explicit language legislation

Our corpus contains 108 pieces (7.2%) of explicit language legislation (Table 4). Primary legislation explicitly concerning language constitutes the smallest part of the corpus, just six pieces of legislation (four for Wales, two for Scotland), all of which falls into the domain of Status, protection and promotion of languages, and legislates for indigenous languages. Welsh was initially protected by the Welsh Language Act 1967,³³ until this was repealed (December 1993) and replaced with the Welsh Language Act 1993. The 1993 Act, amongst other things, established the Welsh Language Board with the aim to promote and facilitate the use of Welsh, and obliged public bodies providing services in Wales to produce Welsh language schemes, setting out how each body will allow for the use of Welsh.³⁴ This was followed by the Welsh Language (Wales) Measure 2011 which gave the language official status in Wales and further legislated for the equality and use of Welsh in Wales.³⁵ In 2012, the National Assembly for Wales (Official Languages) Act³⁶ amended the Government of Wales Act 2006: English and Welsh are named the official languages of the Assembly, to be treated on a basis of equality.

Table 4. Explicit legislation categorization.

		Number	% of section
Jurisdiction	Wales	62	57.4%
	England and Wales	18	16.7%
	Scotland	10	9.3%
	England	7	6.5%
	UK	6	5.6%
	NI	5	4.6%
Legislation type	Primary	6	5.6%
	Secondary	102	94.4%
Language type	Indigenous	87	N/A
	ML	17	N/A
	Community	1	N/A
Domain	Education	44	40.7%
	Status, protection and promotion of languages	28	25.9%
	Registration of births, death and marriages	7	6.5%
	Business	6	5.6%
	Elections and voting	5	4.6%
	Law and crime	4	3.7%
	Media	4	3.7%
	Public health and safety	4	3.7%
	Armed forces	2	1.9%
	Health services	2	1.9%
	Government	1	0.9%
	Nationality and citizenship	1	0.9%

In Scotland, the Gaelic Language (Scotland) Act 2005 established the Bòrd na Gàidhlig, tasked with the promotion of Gaelic language, education and culture, ‘with a view to securing the status of the Gaelic language as an official language of Scotland commanding equal respect to the English language’ (Art.1, Para.3).³⁷ As with the Welsh Act, provisions were established to oblige certain public authorities to produce Gaelic language plans, setting out strategies to promote Gaelic and increase its use in Scotland.³⁸ The extent to which Gaelic is protected by other legislation is, however, unclear. For instance, the Patient Rights (Scotland) Act 2011 states that ‘a relevant NHS body must take account of the particular needs of the persons to whom the Charter is to be made available as to the form of the Charter (for example by making it available in different languages or in Braille)’ (Art.1, Para.10).³⁹ It is not specified whether ‘different languages’ is to be interpreted as ‘different from English’ or ‘different from English and Gaelic’; the number of cases in which this would be applicable, however, is low, given the number of monolingual Gaelic speakers.⁴⁰ No equivalent legislation exists for Scots, a language spoken by c.1.5 million people in Scotland (c.30% of the population) according to the 2011 census,⁴¹ and recognized as a minority language by the European Charter for Regional or Minority Languages (Council of Europe, 2020).⁴² Within the chronological scope of our corpus, Scotland also has primary legislation dedicated to BSL, the British Sign Language (Scotland) Act 2015.⁴³ In April 2022, the BSL Act 2022 recognized BSL as a language of England, Wales and Scotland.⁴⁴ There is no comparable legislation for BSL or Irish Sign Language (ISL) in NI (De Meulder, 2015 provides a summary of the legal recognition of sign languages).⁴⁵

There is no explicit primary legislation for the status and protection of the minoritized indigenous languages for the UK, England or NI. This is not, of course, because of an absence of indigenous languages in England and NI: Cornish, for instance, is not explicitly legislated for in the UK,⁴⁶ nor are schools or other public bodies obliged to provide services in Cornish (Sayers et al., 2019, p. 5). Nevertheless, in 2014 a government press release⁴⁷ announced that the Cornish people and their language were to be recognized as a national minority following the Council of Europe’s Framework Convention for the protection of national minorities,⁴⁸ and in 2015 the UK’s fourth report for the framework contained a ‘Cornish language’ section detailing activities using the language, including a weekly five-minute programme in Cornish on BBC Radio Cornwall.⁴⁹

The corpus contains 102 pieces of explicit secondary legislation. Explicit primary legislation was limited to one domain, Status, protection and promotion of languages. Whilst around one fifth of explicit secondary legislation (n = 22) also concerns this domain, it also crosses various government departments (12 of the 21 domains, including Education, Registration of births, deaths and marriages, and Business), and moves beyond legislating for indigenous languages alone. Regarding language type, over half of the 102 pieces contain stipulations concerning the Welsh language (n = 68, 66.7%), including 12 pieces of legislation whose jurisdiction is both England and Wales. Explicit secondary legislation is categorized in Table 5.

Table 5. Explicit secondary legislation categorization.

		Number	% of section
Jurisdiction	Wales	58	56.9%
	England and Wales	18	17.6%
	Scotland	8	7.8%
	England	7	6.9%
	UK	6	5.9%
	NI	5	4.9%
Language type	Indigenous	81	N/A
	ML	17	N/A
	Community	1	N/A
Domain	Education	44	43.1%
	Status, protection and promotion of languages	22	21.6%
	Registration of births, death and marriages	7	6.9%
	Business	6	5.9%
	Elections and voting	5	4.9%
	Law and crime	4	3.9%
	Media	4	3.9%
	Public health and safety	4	3.9%
	Armed forces	2	2.0%
	Health services	2	2.0%
	Government	1	1.0%
	Nationality and citizenship	1	1.0%

One key difference from the primary legislation is the presence of legislation whose remit is England or UK-wide. Moreover, explicit primary legislation exclusively concerned indigenous languages—still the most frequent language type for secondary legislation (n = 80)—, but secondary legislation also relates to ML education (n = 17), where we find most of the legislation relating to England or the UK more broadly (n = 7). Just one piece of legislation provides for community-language speakers (The Right to Interpretation and Translation in Criminal Proceedings (Scotland) Regulations 2014).⁵⁰ The Law and Crime domain similarly provides legislation explicitly concerning Irish⁵¹—the only example of explicit secondary legislation for Irish in the corpus outside the Education domain.

Two domains dominate explicit secondary legislation: Education (n = 44, 43.1%) and Status, protection and promotion of languages (n = 22, 21.6%). The remaining ten domains contain fewer examples but, nevertheless, testify to the breadth of domains in which explicit language legislation occurs. For instance, seven items (6.9%) concern the Registration of births, deaths and marriages, all of which provide for registrations to be conducted in Welsh as well as, or instead of, in English. Of the 44 examples of Education legislation, 17 concern ML, which perhaps explains the perception that much UK language legislation relates to ML education. ML legislation determines the statutory requirements for language teaching in schools, including the age from and to which languages should be taught (e.g. the Education (National Curriculum) (Modern Foreign Languages) Order 1989),⁵² and student exemptions from languages education (e.g. the Education (National Curriculum) (Exceptions at Key Stage 4) Regulations 1998).⁵³

Whilst educational policy dominates explicit secondary legislation, 58 pieces of legislation (56.9%) occur across eleven other domains. As already noted, further (more limited) legislative protection for sign language(s) is found in Media legislation. The Broadcasting (Sign Language) Order 1997 requires that 5% of programming includes sign language (which sign language(s) is not specified).⁵⁴ This obligation, however, comes with a significant caveat; excluded from the requirement are any programmes ‘which the ITC [Independent Television Commission] considers it to be inappropriate for the obligation to apply’.

In Scotland, Media legislation ensured a platform for the Gaelic language almost 20 years before the Gaelic language (Scotland) Act 2005. The Multiplex Licence (Broadcasting of Programmes in Gaelic) Order 1996,⁵⁵ and the 2008 order amending it,⁵⁶ made provision for Gaelic-language programming. Specifically, the 2008 order obliged S4C Digital Networks (a company which owned six terrestrial television channels) to broadcast at least 30 minutes of Gaelic-language programming between 18:00 and 22:30 daily on certain of its channels. This was revoked in 2011 following a new agreement which would bring BBC ALBA (a Gaelic-language television channel) to Freeview (a free-to-view television platform): consequently, ‘The Secretary of State now considers it unnecessary to require the provision of Gaelic programming’ (The Multiplex Licence (Broadcasting of Programmes in Gaelic) (Revocation) Order 2011).⁵⁷

Included in the large proportion of explicit secondary legislation that concerns Welsh are the Welsh Language Schemes Orders and Welsh Language Standards Regulations which follow the requirements specified in the Welsh Language Act. Much of the legislation involves increasing or ensuring provision for the use of Welsh in administration, e.g. the Welsh Language (Gambling and Licensing Forms) Regulations 2010 allow for documents to be completed in Welsh or in Welsh and English in certain situations.⁵⁸ Such legislation allows Welsh-speakers increasingly to conduct more of their daily life in Welsh and moves Wales closer to a situation of equal bilingual status. Similarly, the Right to Interpretation and Translation in Criminal Proceedings (Scotland) Regulations 2014 oblige courts and the police to assist persons who either do not speak or understand English or have hearing or speech impediments, thus potentially enabling everyone to partake in proceedings.⁵⁹

Finally, we find secondary legislation establishing provisions for effective and safe communication in the Transport domain. For instance, the Merchant Shipping (Minimum Standards of Safety Communications) Regulations 1997 set out provisions intended to ensure the safety of a ship’s crew and passengers,⁶⁰ e.g. Art.5(a) states that there must be ‘means in place for effective oral communication related to safety between all members of the ship’s or hovercraft’s company’ and ‘adequate means of communication’ in English between the ship/hovercraft and onshore authorities. In such cases the legislation is not intended to provide for the status or protection of a language but foregrounds the importance of language as a communicative tool. This example which applies to the UK as a whole demonstrates a still dominant underlying monolingual ideology.

Even in this small amount of explicit language legislation, we see that the domains in which it is enacted reach far beyond the DfE. Although legislation concerning Education constitutes the principal domain (43.1%), eleven other domains are involved. Indigenous languages, particularly Welsh, but also Gaelic, Irish and BSL, are frequently mentioned in explicit legislation, but only one example concerned with community-language speakers was identified.

More hidden

We turn now to legislation primarily conceived as being about a topic other than language, but which contains one or more clauses relating to language(s). Indeed, as the following analysis will demonstrate, it may even mark a milestone in recognizing the status of a language. As noted above, 1,393 pieces of primary and secondary legislation in the corpus contain a hidden language element or elements. This represents 92.8% of the total corpus, a finding which may account for the general perception that the UK has relatively little language policy.

Legislation with a hidden language element is categorized in Table 6. It spans all three language types and 19 of the 21 domains (no hidden legislation relates to the Armed forces or the Status, protection and promotion of languages domain—the latter, by definition, only contains explicit language legislation). Indigenous languages constitute the most frequently occurring language type in hidden legislation (n = 599) as in explicit language legislation, confirming that these are a considerable focus of legislation; with an overwhelming majority, the Welsh language is mentioned the most frequently. Legislation concerning community languages, although less frequent at 269 instances, still considerably outweighs legislation on ML. This is in stark contrast to the one example of explicit language legislation relating to community-language speakers.

Table 6. Hidden legislation categorization.

		Number	% of section
Jurisdiction	UK	577	41.4%
	Scotland	205	14.7%
	Wales	178	12.8%
	NI	174	12.5%
	England	145	10.4%
	England and Wales	110	7.9%
	England, Wales and NI	2	0.1%
	England, Wales and Scotland	2	0.1%
Legislation type	Primary	187	13.4%
	Secondary	1,206	86.6%
Language type	Indigenous	599	N/A
	Community	269	N/A
	ML	73	N/A
Domain	Public health and safety	312	22.4%
	Education	239	17.2%
	Law and crime	191	13.7%
	EU law and Brexit	114	8.2%
	Social care	92	6.6%
	Finance	82	5.9%
	Business	78	5.6%
	Health services	51	3.7%
	Transport	47	3.4%
	Elections and voting	44	3.2%
	Government	22	1.6%
	Media	22	1.6%
	Registration of births, death and marriages	21	1.5%
	Land and property	19	1.4%
	Nationality and citizenship	19	1.4%
	Census	15	1.1%
	Tax	4	0.3%
	Charities	3	0.2%
	Other	17	1.2%

Regarding the primary legislation containing hidden language elements (n = 187, 13.4% of the corpus), 18 of the 21 domains are concerned (those excluded being Armed forces; Tax; Status, protection and promotion of languages) and all three language types. Table 7 details the categorization of this section. The largest amount of primary hidden language legislation concerns the UK as a whole—perhaps unsurprising given that the UK Government has been able to enact primary legislation in all domains since the corpus start date and, post-devolution, the devolved administrations cannot enact primary legislation in the reserved areas. Once again, indigenous languages are the most frequently legislated for (n = 132). Community languages are invoked in 45 examples—more than in explicit language legislation, but still considerably less frequently than indigenous languages. Finally, eight pieces of primary legislation concern ML educational policy; we shall see that this is more frequently legislated for in secondary legislation.

Table 7. Hidden primary legislation categorization.

		Number	% of section
Jurisdiction	UK	97	51.9%
	Scotland	55	29.4%
	Wales	17	9.1%
	NI	14	7.5%
	England and Wales	4	2.1%
Language type	Indigenous	132	N/A
	Community	45	N/A
	ML	8	N/A
Domain	Finance	34	18.2%
	Law and crime	28	15.0%
	Education	27	14.4%
	Social care	13	7.0%
	Land and property	12	6.4%
	Business	11	5.9%
	Elections and voting	11	5.9%
	Government	9	4.8%
	Media	7	3.7%
	Nationality and citizenship	7	3.7%
	Transport	5	2.7%
	Health services	4	2.1%
	Registration of births, death and marriages	4	2.1%
	Census	2	1.1%
	Charities	2	1.1%
	EU law and Brexit	2	1.1%
	Public health and safety	2	1.1%
	Other	7	3.7%

Finance (n = 34, 18.2%), Law and crime (n = 28, 15.0%) and Education (n = 27, 14.4%) are the three domains in which we find the most primary legislation with a hidden language element, together comprising almost half this section. Legislation concerning Finance rather than Education heads the list: this primarily concerns the allocation of government budgets (31 of 34, 91.2%). Pre-devolution, the Appropriation Act 1991,⁶¹ allocated funds to Gaelic and Welsh-language broadcasting. Post-devolution, the Budget (Scotland) Acts (2000–2021) fund the promotion and development of Gaelic, e.g. the 2008 budget assigns funds to support ‘the cultural heritage of Scotland, including the Gaelic language’ (S.1).⁶² In our most recent budget (2021), financial support is provided for the Gaelic language, the Bòrd na Gàidhlig, the Gaelic Media Service 20 (MG Alba), and, for the first time, the Scots language (S.1).⁶³ The remaining three examples of primary Finance legislation concern the language(s) to be used for certain documents, e.g. the Finance Act 2002 stipulates the conditions under which a VAT invoice in a language other than English must be translated into English (Art.24, Para.6).⁶⁴ This type of stipulation is common across hidden legislation, particularly in the Business domain.

As already noted, hidden language legislation often contains landmark provision for languages. In the Nationality and citizenship domain (n = 7, 3.7%), for example, primary legislation sets out the language(s) a person must speak to become a British citizen. The British Nationality Act 1948 states that one requirement for the ‘naturalisation of an alien’ is ‘sufficient knowledge of the English language’ (S.2, 1.d).⁶⁵ This is amended in the Immigration Act 1971 to include Welsh (Appendix A to S.1, Art.5A),⁶⁶ and in the British Nationality Act 1981 Gaelic (S.1, Art.1)⁶⁷—equating, in legislative terms, English, Welsh and Gaelic as legal markers of ‘Britishness’. Given that the Gaelic Language (Scotland) Act was not enacted until 2005, this represented a significant milestone for the status of Gaelic in 1981.⁶⁸ The Immigration Act 2014 serves as a point of contrast here, amending Art.8 of the Nationality, Immigration and Asylum Act 2002 as follows:

(2) It is in the public interest, and in particular in the interests of the economic wellbeing of the United Kingdom, that persons who seek to enter or remain in the United Kingdom are able to speak English, because persons who can speak English—(a) are less of a burden on taxpayers, and (b) are better able to integrate into society.⁶⁹

Whereas earlier Nationality and citizenship legislation was inclusive of certain indigenous languages, the 2014 Act stresses the importance of English, reinforcing a monolingual ideology that positions a monolingual society as practical and inclusive.

Whilst no explicit primary legislation contains stipulations for community languages, 45 examples of primary legislation with hidden language elements relate to community languages and their speakers. Most commonly, this legislation concerns the provision of documents and information in languages other than English. For instance, the Electoral Administration Act 2006 lists the alternative formats in which documents should be made available for voters, including Braille, and ‘languages other than English’.⁷⁰ While not explicitly stated, this presumably encompasses provision for community-language as well as indigenous-language speakers. The decision about the language(s)/format(s) in which to provide information is left to ‘the person who is required or authorized to give or display the document’ whose choice should be made ‘as he thinks appropriate’ (Art.36), potentially limiting the legislation’s effects.

Community-language speakers are more explicitly referenced in the Local Government (Scotland) Act 1966, but again with restrictions (our emphasis), including only benefitting immigrants from the Commonwealth, i.e. a subsection of the UK’s community-language speakers:

the Secretary of State may pay to local authorities who in his opinion are required to make special provision in the exercise of any of their functions in consequence of the presence within their areas of substantial numbers of immigrants from the Commonwealth whose language or customs differ from those of the community (Art.11).⁷¹

In the Local Government in Scotland Act 2003,⁷² these provisions for additional funding for local authorities where there are substantial numbers of community-language speakers, whether from the Commonwealth or elsewhere, no longer feature.

Community-language speakers are also mentioned in primary legislation concerning Education. Found throughout hidden primary and secondary legislation is the idea that speaking a community language is not equivalent to a learning difficulty, e.g. ‘a person is not to be taken to have a learning difficulty solely because the language (or form of language) in which the person is or will be taught is different from a language (or form of language) which has at any time been spoken in the person’s home’ (Art.41).⁷³ The need for such stipulations, of course, raises questions surrounding the treatment of community-language speakers in educational settings. There is no further provision for community languages or community-language speakers in primary Education legislation.

Beyond Education, much hidden primary legislation makes practical stipulations concerned with facilitating access to services and information through language choice and/or language assistance. For instance, the Pensions Appeal Tribunals Act 1943 states that an appellant or witness may use Welsh, ‘if he considers that he would otherwise be at a disadvantage by reason of his natural language of communication being Welsh’.⁷⁴ The Act is not limited to Wales and England (e.g. S.6, Art.6 mentions Scotland and NI), yet no provision is made for any other languages.

Other examples of hidden primary legislation make ideological links between language(s) and culture(s). The revised Wireless Telegraphy Act 2006, for instance, states that wireless telegraphy licences (e.g. radio stations) are granted only if the station ‘is not likely to … prejudice the promotion of cultural and linguistic diversity’ (Art.8).⁷⁵ A similar collocation of ‘cultural and linguistic’ is found in The Islands (Scotland) Act 2018, which offers explicit protection for speakers and acknowledgement of the UK’s linguistic diversity. It requires that Scottish ministers ‘have regard to the … cultural characteristics (including the linguistic heritage) of each of the areas inhabited by island communities’ when preparing the national islands plan (Art.4).⁷⁶

Hidden language legislation is also used, quite explicitly, to hierarchize languages in cases where multiple language versions of documents exist. The Merchant Shipping and Maritime Security Act 1997 establishes equal footing for the multiple language versions of the convention: ‘This Convention is established in a single original in the Arabic, Chinese, English, French, Russian and Spanish languages, each text being equally authentic’ (Art.54).⁷⁷ In other cases, one language version is given precedence. Usually, this is English; e.g. the Criminal Procedure (Amendment) Rules 2009 state that ‘In the case of a discrepancy between an English and Welsh text the English text shall prevail’ (S.1).⁷⁸ However, there are examples where English must concede to other languages, e.g. The Carriage by Air Act 1961, where, in cases of inconsistencies between the English text in Part I of S.1 and the French text in Part II, ‘the text in French shall prevail’ (Art.1).⁷⁹ Whilst the establishment of linguistic precedence may be necessary in cases of discrepancy, the choice of language and the implementation of a linguistic hierarchy cannot be divorced from ideology.

Secondary legislation with a hidden language element constitutes the largest part of the corpus (n = 1,206, 80.3%). Its categorization is presented in Table 8. UK-wide legislation dominates (n = 480), while legislation relating to one of the four constituent jurisdictions is broadly equal in number. This differs from explicit legislation where legislation relating to Wales predominated. Quantitative analysis of the relevant legislation across the four constituent jurisdictions again demonstrates the breadth of domains in which language legislation is enacted, with 19 represented. Four domains dominate: Education, Social care, Public health and safety, and Law and crime.

Table 8. Hidden secondary legislation categorization.

		Number	% of section
Jurisdiction	UK	480	39.8%
	Wales	161	13.3%
	NI	160	13.3%
	Scotland	150	12.4%
	England	145	12.0%
	England and Wales	106	8.8%
	England, Wales and NI	2	0.2%
	England, Wales and Scotland	2	0.2%
Language type	Indigenous	467	N/A
	Community	224	N/A
	ML	65	N/A
Domain	Public health and safety	310	25.7%
	Education	212	17.6%
	Law and crime	165	13.7%
	EU law and Brexit	112	9.3%
	Social care	79	6.6%
	Business	67	5.6%
	Finance	48	4.0%
	Health services	47	3.9%
	Transport	42	3.5%
	Elections and voting	33	2.7%
	Registration of births, death and marriages	17	1.4%
	Media	15	1.2%
	Census	13	1.1%
	Government	13	1.1%
	Nationality and citizenship	12	1.0%
	Land and property	7	0.6%
	Tax	4	0.3%
	Charities	1	0.1%
	Other	9	0.7%

A large amount of hidden secondary legislation specifies the language(s) in which documents are to be provided and details of how and when these are to be translated. The Public health and safety domain (n = 310, 25.7%) frequently concerns language choice because of the importance of information being widely and readily understood. English is often given precedence, e.g. the Explosives (Fireworks) Regulations (Northern Ireland) 1999 state that ‘A person shall not sell or transfer any firework which is not labelled in English’ (Art.8).⁸⁰ The language(s) to be used are not always explicitly listed; for instance, the Food Labelling Regulations 1996 state that product labels much be ‘in a language easily understood by the purchaser’ (Art.3),⁸¹ and the Radioactive Material (Road Transport) (Great Britain) Regulations 1996 specify that statements for carriers of radioactive material ‘shall be in the languages deemed necessary by the carrier or the competent authorities concerned’ (S.21), allowing for multiple language versions. Since these examples often concern trade, the choice of language(s) depends on the product’s country of origin/destination. Other indigenous languages are also mentioned; for instance, the Scotch Whisky Regulations 2009 stipulate that notices ‘must be in English and Welsh’ (Art.41),⁸² whilst the Traffic Signs Regulations and General Directions 2016 include guidelines for the use of Gaelic ‘in addition to, or instead of’ English on certain signs in Scotland (e.g. Art.38).⁸³ These labelling conventions seem to suggest that they are based on the particular context and pragmatic considerations about what is needed to guarantee the public’s health and safety. Nevertheless, the minimal inclusion of languages other than English seems again to strongly reflect a monolingual ideology.

The second largest domain here is Education and the legislation primarily relates—perhaps predictably—to the teaching of indigenous and ML, e.g. the Education (Amendment of the Curriculum Requirements for Fourth Key Stage) (England) Order 2003 which made ML education optional at KS4,⁸⁴ or the Education (Special Schools) Regulations 1994 which, among other things, established guidelines for Welsh education in Wales.⁸⁵ However, we also find provisions for community-language speakers via funding for EAL teaching, e.g. the School and Early Years Finance (England) Regulations 2021 set out funding provisions for pupils whose L1 is not English.⁸⁶ Similar funding has been available in England since at least 1995. Other secondary legislation concerning pupil and school performance information requires schools to report the number of pupils whose first language is not English, e.g. the Education (School Performance Information) (England) Regulations 2007,⁸⁷ and information concerning educational and welfare provision for pupils whose L1 is not English (The Education (Independent School Standards) (England) Regulations 2010, Part 6).⁸⁸ Significantly, all Education legislation concerning community-language speakers relates to the teaching of English: no mention is made of teaching or promoting pupils’ L1 or other languages, indicating that the support and teaching of community languages is largely limited to complementary schools.

Recording the language(s) spoken in society is a concern in various domains. An obvious case is the census, legislated for, to cite a recent example, in the Census (England) Regulations 2020.⁸⁹ The census asks ‘What is your main language?’ and allows the naming of just one additional language; a highly problematic question (Sebba & Ayres-Bennett, 2021), not least in implying that those speaking more than one language have a ‘main’ language and in denying the possibility of multilingualism. Conversely, in the Health services domain, we find provision for the recording of multilingualism amongst healthcare professionals, e.g. the National Health Service (General Dental Services) Regulations 1992 state: ‘The dental list may … give particulars of any languages, other than English, spoken by the dentist’ (Art.4).⁹⁰

Secondary legislation in the Media domain again makes provision for indigenous languages not found elsewhere in the corpus. The Cultural Test (Television Programmes) Regulations 2013,⁹¹ and the Cultural Test (Video Games) Regulations 2014,⁹² outline points-based ‘cultural tests’, used to determine whether a programme or game can be certified as ‘British’ and consequently benefit from tax relief. Points are gained for original dialogue recorded in English or ‘a recognized regional or minority language’, where ‘“recognised regional or minority language” means Welsh, Scottish-Gaelic, Irish, Scots, Ulster Scots, Cornish or British Sign Language’. Thus, indigenous languages, not protected or promoted in other legislation (e.g. Cornish), are explicitly promoted in cultural products. The regulations for films differ slightly; in the Films (Definition of “British Film”) Order 2006,⁹³ a film earns points for dialogue in English or ‘a recognised regional or minority language’, although here the definition excludes BSL. The regulations for television programmes and films are amended by the Cultural Test (Television Programmes) (Amendment) Regulations 2015,⁹⁴ and The Films (Definition of “British Film”) Order 2015 respectively;⁹⁵ ‘the English language or a recognised regional or minority language’ is replaced by ‘a language recognised for official purposes in the United Kingdom or another EEA state’, meaning that points will no longer be awarded for dialogues in most of the previously listed indigenous languages.

Concerning Law and crime, provisions are established to ensure that those involved in proceedings understand and are understood. For instance, the Juvenile Justice Centre Rules (Northern Ireland) 2008 state: ‘Where a child is not fluent in English all reasonable steps shall be taken to provide the services of an interpreter’ (Art.11).⁹⁶ In some cases, this extends to include English-speakers with communication difficulties, e.g. the Residential Property Tribunal Procedure (England) Regulations 2006 state that participants who are ‘unable to read or speak or understand the English language’ or are ‘without hearing or speech’ must be provided with free assistance (Art.36).⁹⁷ Some Scottish provision seems broader still. For example, the Mental Health Tribunal for Scotland (Practice and Procedure) Rules 2005 specify that ‘the Tribunal shall take all reasonable steps’ to assist a person who ‘has difficulty in communicating or generally communicates in a language other than English’ (Art.53).⁹⁸ Whilst ‘all reasonable steps’ clearly mitigates the obligation, the range of participants covered can potentially be interpreted more broadly than in the English legislation. It is also interesting to see speaking ‘a language other than English’ coordinated with ‘difficulty in communicating’, the two requiring rather different forms of support.

Language is not a protected characteristic in the UK, unlike, for instance, race or religion,⁹⁹ yet in certain secondary legislation it receives the same protections (or promises of protection), e.g. The Prisons and Young Offenders Institutions (Scotland) Rules 1994:

the Governor shall seek to eliminate within the prison discrimination on the grounds of gender, sexual orientation, race, colour, language, religion, political or other opinion, national or social origin, association with a national minority, birth, medical condition and economic or other status against particular prisoners or categories of prisoners (Art.4).¹⁰⁰

Protections against linguistic discrimination are also found in Health services and Social care secondary legislation, e.g. Art.16 of the Private and Voluntary Health Care (England) Regulations 2001¹⁰¹ states that suitable arrangements shall be made to ensure that a healthcare establishment is conducted ‘with due regard to the sex, religious and spiritual needs, racial origin, and cultural and linguistic background and any disability of patients’.

Conclusion

Our results help explain certain misconceptions about UK language policy. First, the perception that there is little or no UK language policy is countered by the fact that our corpus comprises c.1,500 pieces of legislation, covering all three language types, and 21 different domains. However, over 90% of it is hidden in legislation primarily about another issue and is therefore generally not very salient either to the public or indeed policymakers, even though it may sometimes contain landmark provisions.

A second misconception is that UK language policy principally concerns ML. Overall indigenous languages dominate, particularly in explicit legislation concerning their status, protection and promotion; since, however, most of this concerns Welsh and Gaelic, it is perhaps overlooked by those focussed on Whitehall, where the explicit secondary legislation does indeed concentrate on Education, of which ML education forms a significant proportion. Moreover, despite the predominance of legislation relating to indigenous languages, much of this relates to Welsh and, to a lesser extent, Gaelic; Scots, Cornish and Manx feature minimally in legislation in other domains, and legislation for Irish remains highly contentious. What was, however, confirmed was the relatively little legislation relating to community languages, and the vast majority of what does exist is hidden and may be coloured by a prevailing monolingual ideology.

A third misconception is that language legislation is concentrated in the Education domain. Overall, 21 domains were represented, the top three being Public health and safety, followed by Education and Law and crime. Where public safety or access to justice are concerned, practical considerations outweigh the impetus to promote English over other languages.¹⁰² We have found abundant evidence that language issues permeate society, from prisons, appeals and tribunals to health and social care, food labelling and the media, not just the obvious domains concerning corpus and acquisition planning.

These findings have clear implications for language policy, not only in the UK, but more widely. When examining countries which are believed to have strong and explicit language policy, such as France, it is likely that there will also be a portfolio of hidden language legislation that needs to be explored to avoid gaining an incomplete picture of the position of languages there. In the UK, the fact that the vast majority of language legislation is hidden militates against having a coherent joined-up strategy. Since portfolios of legislation are frequently conceived as dealing with another topic, the language dimension may be minimized or overlooked, meaning that language legislation does not receive the attention it deserves. Moreover, legislation is prepared and written by civil servants and policymakers who are primarily specialists in other domains and who may not therefore appreciate the significance of the language dimension or have little or no training in preparing legislation relating to languages. This variance is reflected in the lack of consistency of phrasing in different domains of legislation and suggests that civil servants might benefit from training which raised awareness of how their portfolios might have a language(s) dimension.

Our analysis of a corpus of legislation from across the four UK jurisdictions has furthermore highlighted the importance of cross-jurisdictional comparison and suggested that this is also a fertile approach for future research. Devolution has given each jurisdiction fresh power and impetus to legislate according to its own priorities. In the absence of an overarching UK language policy, this delegated authority is resulting in increasingly divergent language policy across the UK. This is evidenced, for instance, in Scotland’s initiative to legislate for BSL in the British Sign Language (Scotland) Act 2015, or by the sheer amount of explicit language legislation coming from Wales (especially since the Welsh Language (Wales) Measure 2011), in comparison to other jurisdictions. Whilst provision for languages education was already different pre-devolution, here too we see increased divergence, with Scotland implementing its 1+2 language policy, and Wales promoting bilingualism + 1, as part of its policy to reach a million Welsh speakers by 2051.

Our analysis has not addressed the thorny question of the extent to which the legislation we have identified has been implemented in full or in part, and this is clearly a potentially fruitful area for future research. We noted that the situation for indigenous language provision can be unclear. For instance, whilst the Gaelic language is protected by the Gaelic Language (Scotland) Act 2005 in Scotland, it is unclear whether or when Gaelic is protected by other legislation applicable to the jurisdiction, as was illustrated above with the Patient Rights (Scotland) Act 2011 where it is unclear whether the reference to ‘different languages’ means different from English or different from English and Gaelic. Such ambiguity in the legislation leaves indigenous language provision open to being overlooked. We have, moreover, noted that legislation may use permissive modal auxiliaries or hedging clauses, reducing notably the force of the legislation.

For language policy in the UK to move higher up the political agenda and to counter the present patchiness of the legal framework, greater cross-government collaboration—across ministries, departments and jurisdictions—is needed. Those drafting legislation might also be encouraged to consider more systematically whether there is a language dimension to their portfolio, just as they now automatically consider whether there is a gender or ethnic dimension to their work. We hope that highlighting the breadth of language legislation which already exists will constitute a positive first step in raising awareness not just about what exists, but what remains to be done.

Acknowledgements

This study was generously funded by the UK’s Arts and Humanities Research Council grant, Promoting Language Policy (AH/V004182/1). We would also like to thank the Legislative Codes Office within the Office of the Legislative Counsel (Welsh Government), Louise Glen (Education Scotland) and Linsey Farrell (NI Department of Education) for their help in compiling the corpus and the two anonymous peer reviewers for their insightful feedback. All errors and omissions remain our own.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Arts and Humanities Research Council: Grant Number AH/V004182/1.

Notes on contributors

Emma Humphries

This research was conducted whilst Emma Humphries was the Research Project Manager of the AHRC-funded Promoting Language Policy project. She completed her PhD in French sociolinguistics at the University of Nottingham in January 2021 and she is currently a Research Fellow on the AHRC-funded project ‘Foreign, indigenous and community languages in the devolved regions of the UK: policy and practice for growth’, led by Professor Janice Carruthers at Queen’s University Belfast. Her research interests include language attitudes and ideologies, language policy and computer-mediated communication.

Wendy Ayres-Bennett

Wendy Ayres-Bennett is Emerita Professor of French Philology and Linguistics, University of Cambridge, and was PI on the Promoting Language Policy research project, funded by the AHRC (2020-2022). She has published widely on the history of French and the history of linguistic thought. Her current research interests include language policy, standardization, codification and ideologies. Between 2016 and 2021, she was Principal Investigator on the multi-disciplinary, multi-institution research project, Multilingualism: Empowering Individuals, Transforming Societies, funded by the AHRC under its Open World Research Initiative (www.meits.org). The project worked closely with policymakers and practitioners to promote the value of languages for key issues of our time and the benefits of language learning for individuals and societies.

Notes

1 https://www.legislation.gov.uk/ukpga/1993/38/enacted/data.pdf. All online resources were last accessed 18/1/2022.

2 https://www.legislation.gov.uk/asp/2005/7/enacted/data.pdf.

3 https://www.legislation.gov.uk/uksi/1995/57/made/data.pdf.

4 Education (School Performance Information) (England) (Amendment) Regulations 2009: https://www.legislation.gov.uk/uksi/2009/646/pdfs/uksi_20090646_en.pdf.

5 https://www.legislation.gov.uk/ukpga/1998/47/enacted/data.pdf.

6 https://www.legislation.gov.uk/ukpga/1998/46/enacted/data.pdf.

7 https://www.legislation.gov.uk/ukpga/1998/38/enacted/data.pdf.

8 https://www.legislation.gov.uk/ukpga/2012/11/enacted/data.pdf.

9 https://www.legislation.gov.uk/ukpga/2016/11/enacted/data.pdf.

10 On Welsh devolution, see https://senedd.wales/how-we-work/history-of-devolution/.

11 https://www.legislation.gov.uk/ukpga/2006/32/data.pdf.

12 For details, see https://commonslibrary.parliament.uk/senedd-cymru-why-has-the-national-assembly-for-wales-changed-its-name/.

13 In this section we refer to language education policy more broadly, whereas the corpus used for our analysis as described in the following sections is restricted to policy enshrined in primary and secondary legislation.

14 On NI education, see https://www.qub.ac.uk/sites/irishhistorylive/IrishHistoryResources/Shortarticlesandencyclopaediaentries/Encyclopaedia/LengthyEntries/Education/.

15 These are GCSEs taken in England, NI and Wales at the end of KS4 and A Levels taken at the end of KS5. The system differs in Scotland (see https://scqf.org.uk/about-the-framework/interactive-framework/).

16 Examination entries have similarly fallen dramatically in England and NI (Tinsley & Doležal, 2018; Henderson & Carruthers, 2021).

17 https://hwb.gov.wales/curriculum-for-wales/.

18 https://www.legislation.gov.uk/understanding-legislation.

19 https://www.legislation.gov.uk/understanding-legislation.

20 https://www.promotinglanguagepolicy.org/.

21 https://www.legislation.gov.uk/. For more information, see https://www.legislation.gov.uk/understanding-legislation#Whatlegislationisheldonlegislationgovuk.

22 https://www.legislation.gov.uk/ukpga/Geo5/8-9/39/contents/enacted.

23 https://www.legislation.gov.uk/uksi/1996/1350/made/data.pdf.

24 https://www.legislation.gov.uk/uksi/1992/3092/made/data.pdf.

25 https://www.legislation.gov.uk/uksi/1989/1275/made/data.pdf.

26 https://www.legislation.gov.uk/uksi/1993/252/made/data.pdf.

27 https://www.legislation.gov.uk/uksi/2010/245/made/data.pdf.

28 In our categorization, ML is understood in the narrow sense of ML educational policy, typically in the UK the teaching of French, German, Spanish and, to a lesser extent, Italian, Mandarin, Russian, etc.

29 https://www.legislation.gov.uk/uksi/1990/1599/made/data.pdf.

30 https://www.legislation.gov.uk/ukpga/1993/45/pdfs/ukpga_19930045_en.pdf. Throughout we use Art. for Article, Para. for Paragraph and S. for Schedule.

31 132 pieces of legislation relate to multiple jurisdictions.

32 All percentages are rounded to 1dp.

33 https://www.legislation.gov.uk/ukpga/1967/66/enacted/data.pdf.

34 See, e.g., Ofcom’s scheme: https://www.ofcom.org.uk/__data/assets/pdf_file/0014/10481/welsh-language-scheme.pdf.

35 https://www.legislation.gov.uk/mwa/2011/1/enacted/data.pdf.

36 https://www.legislation.gov.uk/anaw/2012/1/data.pdf.

37 https://www.legislation.gov.uk/asp/2005/7/enacted/data.pdf.

38 See, e.g., the Scottish Parliament’s plan: https://archive2021.parliament.scot/Gaelic/GLP201818English.pdf.

39 https://www.legislation.gov.uk/asp/2011/5/enacted/data.pdf.

40 In the 2011 census, c.57,000 people identified as Gaelic speakers: https://www.scotlandscensus.gov.uk/census-results/at-a-glance/languages/.

41 Note that there are several difficulties surrounding the collection of language data in the 2011 census (Sebba, 2017; Sebba & Ayres-Bennett, 2021).

42 In 2020, Scottish association Oor Voyce launched a campaign for the legal recognition of Scots via a Scots Language Act and the establishment of a board comparable to Bòrd na Gàidhlig. Creative Scotland, a non-departmental public body of the Scottish Government, published an advisory Scots Language Policy in 2015 which aimed to promote the use and learning of Scots and enhance its status.

43 https://www.legislation.gov.uk/asp/2015/11/enacted/data.pdf.

44 https://bda.org.uk/bsl-act-now/.

45 As we shall see in the next section, more restrictive legislation provides some legal protections for BSL users in the UK.

46 Cornish does feature in secondary legislation concerning the taxation of cultural products (see hidden legislation section).

47 https://www.gov.uk/government/news/cornish-granted-minority-status-within-the-uk.

48 https://www.coe.int/en/web/conventions/full-list/-/conventions/rms/090000168007cdac.

49 https://rm.coe.int/CoERMPublicCommonSearchServices/DisplayDCTMContent?documentId=09000016805a8c52.

50 ‘Interpretation assistance’ is defined in the legislation as follows: ‘in relation to a person who does not speak or understand English, an oral translation of—(i) the police proceedings or criminal proceedings into the person’s native language or any other language which the person speaks or understands’: https://www.legislation.gov.uk/ssi/2014/95/made/data.pdf.

51 https://www.legislation.gov.uk/nisr/2019/138/made/data.pdf.

52 For a list of those languages termed ‘unconditionally specified as foundation subjects’ (S.1) as well as those which may be considered foundation subjects (S.2) providing the school also offers one or more of the languages in S.1, see https://www.legislation.gov.uk/uksi/1989/825/made/data.pdf.

53 https://www.legislation.gov.uk/uksi/1998/2021/made/data.pdf.

54 https://www.legislation.gov.uk/uksi/1997/167/made/data.pdf.

55 https://www.legislation.gov.uk/uksi/1996/2758/made/data.pdf.

56 https://www.legislation.gov.uk/uksi/2008/1421/pdfs/uksi_20081421_en.pdf.

57 https://www.legislation.gov.uk/uksi/2011/1169/made/data.pdf. Further provision for Gaelic is found in secondary legislation concerning Education (e.g. the Grants for Gaelic Language Education (Scotland) Regulations 1986: https://www.legislation.gov.uk/uksi/1986/410/pdfs/uksi_19860410_en.pdf).

58 https://www.legislation.gov.uk/uksi/2010/2440/made/data.pdf.

59 https://www.legislation.gov.uk/ssi/2014/95/made/data.pdf.

60 https://www.legislation.gov.uk/uksi/1997/529/made/data.pdf.

61 https://www.legislation.gov.uk/ukpga/1991/32/enacted/data.pdf. There is similar legislation 1992-1999.

62 https://www.legislation.gov.uk/asp/2008/2/pdfs/asp_20080002_en.pdf.

63 https://www.legislation.gov.uk/asp/2021/8/enacted/data.pdf.

64 https://www.legislation.gov.uk/ukpga/2002/23/enacted/data.pdf.

65 https://www.legislation.gov.uk/ukpga/Geo6/11-12/56/enacted/data.pdf.

66 https://www.legislation.gov.uk/ukpga/1971/77/enacted/data.pdf.

67 https://www.legislation.gov.uk/ukpga/1981/61/enacted/data.pdf.

68 The Borders, Citizenship and Immigration Act 2009 cites the same languages; this adds that a candidate for naturalization must have ‘sufficient knowledge about life in the United Kingdom’ (Art.40): https://www.legislation.gov.uk/ukpga/2009/11/enacted/data.pdf.

69 https://www.legislation.gov.uk/ukpga/2014/22/enacted/data.pdf.

70 https://www.legislation.gov.uk/ukpga/2006/22/pdfs/ukpga_20060022_en.pdf.

71 https://www.legislation.gov.uk/ukpga/1966/51/pdfs/ukpga_19660051_en.pdf.

72 https://www.legislation.gov.uk/asp/2003/1/enacted/data.pdf.

73 https://www.legislation.gov.uk/ukpga/2009/22/pdfs/ukpga_20090022_en.pdf.

74 https://www.legislation.gov.uk/ukpga/Geo6/6-7/39/enacted/data.pdf.

75 https://www.legislation.gov.uk/ukpga/2006/36/data.pdf.

76 https://www.legislation.gov.uk/asp/2018/12/enacted/data.pdf.

77 https://www.legislation.gov.uk/ukpga/1997/28/pdfs/ukpga_19970028_en.pdf.

78 https://www.legislation.gov.uk/uksi/2009/2087/made/data.pdf.

79 https://www.legislation.gov.uk/ukpga/Eliz2/9-10/27/enacted/data.pdf.

80 https://www.legislation.gov.uk/nisr/1999/392/made/data.pdf.

81 https://www.legislation.gov.uk/uksi/1996/1499/made/data.pdf.

82 https://www.legislation.gov.uk/uksi/2009/2890/made/data.pdf.

83 https://www.legislation.gov.uk/uksi/2016/362/made/data.pdf.

84 https://www.legislation.gov.uk/uksi/2003/2946/made/data.pdf.

85 https://www.legislation.gov.uk/uksi/1994/652/made/data.pdf.

86 https://www.legislation.gov.uk/uksi/2021/59/contents/made.

87 https://www.legislation.gov.uk/uksi/2007/2324/made/data.pdf.

88 https://www.legislation.gov.uk/uksi/2010/1997/made/data.pdf. There is an interesting shift in the wording, from asking for ‘the pupil’s first language’, implying an interest in knowing the language(s) spoken by pupils, to ‘whether English is not the child’s first language’, focusing more on knowledge of English.

89 https://www.legislation.gov.uk/uksi/2020/560/made/data.pdf.

90 https://www.legislation.gov.uk/uksi/1992/661/made/data.pdf.

91 https://www.legislation.gov.uk/uksi/2013/1831/made/data.pdf.

92 https://www.legislation.gov.uk/uksi/2014/1958/made/data.pdf.

93 https://www.legislation.gov.uk/uksi/2006/643/made/data.pdf.

94 https://www.legislation.gov.uk/uksi/2015/1449/pdfs/uksi_20151449_en.pdf.

95 https://www.legislation.gov.uk/uksi/2015/86/pdfs/uksi_20150086_en.pdf.

96 https://www.legislation.gov.uk/nisr/2008/427/made/data.pdf.

97 https://www.legislation.gov.uk/uksi/2006/831/made/data.pdf.

98 https://www.legislation.gov.uk/ssi/2005/420/made/data.pdf.

99 For the full list, see https://www.legislation.gov.uk/ukpga/2010/15/section/4.

100 https://www.legislation.gov.uk/uksi/1994/1931/made/data.pdf.

101 https://www.legislation.gov.uk/uksi/2001/3968/made/data.pdf.

102 Similar conclusions concerning health and safety are reached by Sharma (2018, p. 66).