Language learning guides

People who do not speak English fluently, and people who are d/Deaf, often need an interpreter when they use GP services. NHS England published “Guidance for commissioners: Interpreting and Translation Services in Primary Care” in September 2018. It sets eight principles for the organisations that buy these services in England. This article summarises the principles that matter most to patients, learners and would-be interpreters. The document is guidance rather than law, it was written with GP surgeries in mind, and it does not cover Scotland, Wales or Northern Ireland. It is general information, not a statement of any individual patient’s entitlement.

Free interpreting that the practice arranges

The first principle says interpretation and translation should be free at the point of delivery, of high quality and responsive to a patient’s linguistic needs. Patients should not be asked to pay for interpreting or to provide their own interpreter. Under the second principle, where an interpreter is needed, the primary care provider is responsible for booking one, and it is good practice to confirm in advance the name and gender of the interpreter, if known. The guidance says extra time is needed for a consultation with an interpreter, typically double a regular appointment, and that patients should not wait unnecessarily longer for an appointment because an interpreter is needed, while recognising that availability can affect access.

Preferences and types of interpreting

Patients should be asked about their language requirements and communication needs when they register, and the answers should be recorded and flagged. These can include language, communication needs, the gender of the interpreter and, where relevant, cultural identity. Patients should be told about the types of interpreting available, such as face-to-face, telephone and video remote interpreting or video relay for sign language. Where continuity of care matters, the guidance says systems should allow the same interpreter to be used where practicable. Where a patient has a carer, the carer should be able to access language support, with the patient’s consent.

Relatives, friends and children

The guidance says patients should always be offered a registered interpreter and that relying on family, friends or unqualified interpreters is strongly discouraged and not good practice. If a patient expressly wants a family member or friend to interpret, the patient should give informed consent in their own language, sought independently of that person, and the consent should be noted in the record. Using anyone under 16 for interpretation is not acceptable in any circumstance other than when immediate and necessary treatment is required, and the guidance says this should be exceptional because of safeguarding and competency concerns.

Bilingual staff may help with booking appointments or identifying communication needs, but should not act as interpreters unless that is part of their defined job role and they are qualified, other than for immediate and necessary treatment.

What interpreters are expected to do

Interpreters should be registered with an appropriate regulator, hold suitable qualifications and have the skills to work in health settings. The guidance expects Disclosure and Barring Service checks, annual safeguarding and information governance training, and conduct in a place where the conversation cannot be overheard. Interpreters should introduce themselves and explain their role at the start, interpret for everyone in the appointment, and declare any conflict of interest. They are present only to facilitate communication, with no physical contact or help with clinical procedures, and the guidance notes that interpreters do not provide advocacy.

Qualifications named in the guidance

Annex 2 says spoken-language interpreters should be on the National Register of Public Service Interpreters and hold a Diploma in Public Sector Interpreting (Health), with limited alternatives that include a minimum of NVQ level 3 in interpreting alongside native-level skills or an IELTS 7.5 in English. For British Sign Language, interpreters should be registered with the National Registers of Communication Professionals working with Deaf and Deafblind People, whose requirements include BSL level 6 or an honours degree in a second language. Translators of written documents should hold, among other options, an honours degree in the language or translation or a level 7 translation qualification. The registers are explained in the guides to NRPSI and NRCPD, and career routes in the guide to becoming a translator or interpreter.

The legal position and feedback

The guidance distinguishes two groups. For disabled people, including d/Deaf, blind and deafblind patients, the Equality Act 2010 places a duty on service providers to make reasonable adjustments, and the Accessible Information Standard, in full force from 1 August 2016, requires NHS providers to identify, record, flag, share and meet communication needs. For people who do not speak English, the guidance says this is not a protected characteristic under the Equality Act, but section 13G of the National Health Service Act 2006 requires NHS England to have regard to reducing inequalities in access to health services and in outcomes. Patients and clinicians should be able to give feedback on an interpreting service in their first or preferred language and format.

The bottom line

NHS England’s 2018 commissioning guidance says GP interpreting should be free, arranged by the practice and carried out by a registered interpreter, with relatives and friends a last resort that needs the patient’s informed consent and children under 16 used only for immediate and necessary treatment. It names NRPSI and DPSI (Health) for spoken languages and NRCPD for BSL as the benchmark qualifications. The document is guidance for commissioners in England, not a legal entitlement, and local arrangements vary.

Sources

• Guidance for commissioners: Interpreting and Translation Services in Primary Care, NHS England, September 2018 (england.nhs.uk)

A level French, German and Spanish were designed for students learning a language as an additional language, yet some students bring a home language into the classroom. In England, the exams regulator Ofqual studied this in summer 2016 and published “Native speakers in A level modern foreign languages” in April 2017, in response to stakeholder concerns about grading. This article summarises that report. The research is now several years old, the sample was small, and the authors describe their figures as estimates. It is general information about how the system worked at the time and does not describe current grade boundaries.

Who may enter A level languages

Ofqual’s background section says A level modern foreign languages (MFL) were originally conceived as second-language qualifications. It also states that the exam boards set no formal pre-requisites beyond a recommendation that students have skills equivalent to GCSE, and that there are no limits on who can enter. The qualifications are therefore taken both by students learning the subject as an additional language and by students with experience of the language outside study. Individual schools and colleges may set their own entry requirements. The broader picture for home-language learners is in the guide to heritage language learners.

Why Ofqual looked at the question

The research was prompted by concerns that the proportion of native speakers sitting A level MFL was rising, and that grading methods could disadvantage non-native speakers, because grade boundaries are guided by statistical predictions based on prior attainment, while native speakers might not be rewarded appropriately. The report notes that entries for A level French and German fell by around a third between 2007 and 2016, around 7,000 fewer students, while Spanish and other languages rose. Current entry levels appear in the guide to provisional 2026 entries.

The difficulty of counting native speakers

Ofqual says identifying native speakers is far from straightforward. Native speakers exist on a continuum, from a student raised in France until 16 to one who spends part of each year with a French family or has a French background but has lived in the UK for years, yet the research needed a yes or no answer. It rejected a dictionary definition, because terms such as first language and mother tongue are problematic, and found the National Pupil Database insufficient, since it records only a first language.

Ofqual collected two kinds of evidence: a student questionnaire adapted from a bilingualism tool and teachers’ reports. The student data could not be used, because the questionnaire did not work as intended for students whose only exposure was A level study and some responses were contradictory. The research relied on teachers, and the report warns that teachers may be less familiar with some students’ family background or may assume a highly proficient student has native-speaker characteristics.

What the data showed

Fewer than 28% of schools responded, and the smaller entries for Italian and Russian meant analysis focused on French, German and Spanish. According to teacher reports, native speakers made up 8.7% of French students in the sample, 17.4% of German and 10.1% of Spanish. Among students matched to prior attainment, whose results feed the statistical predictions, the figures were 7.6%, 10.5% and 4.9%. Ofqual concluded this suggests that any effect on standards might be smaller than stakeholders anticipated.

Native speakers outperformed non-native speakers on the overall A level after controlling for prior attainment, gender and school type. The effect was greatest in German and for students with lower prior attainment. Native speakers also outperformed a prior-attainment-based prediction at grade A, particularly in German. The effect of removing them from the predictions was relatively small in French and Spanish, in some cases smaller than the tolerances applied by exam boards.

Conclusions and limits

Ofqual concluded that native and non-native speakers perform differently and that there could be implications for maintaining standards, but that quantifying them was challenging. Contextual information suggested the proportion of native speakers was likely to have increased over time, though the research could not show this. On balance it suggested a small but important effect and recommended considering whether an adjustment to standards was appropriate. The authors added that routinely monitoring native speakers each year would not be possible or proportionate.

The findings come from a 2016 sample and are not evidence about current cohorts. How GCSE grading in French and German was later aligned with Spanish is covered in the guide to GCSE grading adjustments, and the A level content in the guide to A level modern languages.

The bottom line

Ofqual’s 2017 report found that native speakers made up roughly 5% to 17% of the A level language students it could analyse in 2016, depending on subject and sample, and that they tended to outperform non-native peers with similar prior attainment, most clearly in German. The authors stressed that defining and identifying native speakers is hard, that the response rate was low, and that the figures are estimates. Anyone choosing an A level language should treat the report as a snapshot of how standards were discussed then.

Sources

• Native speakers in A level modern foreign languages, Ofqual, April 2017 (gov.uk)

People who speak or study languages sometimes consider teaching them in secondary schools. In England, the Department for Education (DfE) pays bursaries and scholarships to trainees on certain teacher training courses, and languages are among the subjects included. This article summarises the DfE’s funding guidance for initial teacher training for the 2026 to 2027 academic year and its bursary funding manual for the same year. It applies to teacher training in England that leads to qualified teacher status (QTS), and the figures can change from year to year. It is general information, and course providers confirm eligibility in writing.

The amounts for 2026 to 2027

For postgraduate courses in 2026 to 2027, the DfE offers a tax-free bursary of £20,000 for languages, including ancient languages. A scholarship of £22,000 is available for French, German and Spanish only, and no scholarship exists for other languages. The manual says the bursary depends on the subject the trainee trains to teach, not the subject of their degree, and that the award is fixed for the duration of the course.

The amounts are lower than the year before. The guidance for 2025 to 2026 listed £26,000 for languages, including ancient languages, and £28,000 for French, German and Spanish scholarships. The manual warns that no assumptions should be made about the value or scope of awards in future years.

Who is eligible

Applicants may be eligible for a bursary with a 1st, 2:1, 2:2 or master’s degree or a PhD. Scholarships are usually available to applicants with a 1st, 2:1 or master’s degree or a PhD, though in exceptional circumstances they may be awarded to a graduate with a 2:2 and significant relevant experience. Trainees are only eligible if they are entitled to support under the student finance criteria, with an exception for languages and physics trainees who received an offer before 5pm on 7 May 2026. A degree from outside the UK must have the breadth and depth expected of a UK first degree with at least second class honours.

The course must lead to a recommendation for QTS and be delivered by an accredited provider. Courses leading only to an academic award, such as a postgraduate certificate in education without QTS, or to QTLS, do not qualify. A course combining subjects must be 50% or more languages to attract the languages bursary; the manual gives the example of a 50% languages and 50% history course, which receives the languages bursary.

Scholarships

Scholarships are applied for separately from the teacher training application, to the relevant professional body, and are awarded only to applicants who pass an application and interview process. A trainee cannot receive both a scholarship and a bursary, although a trainee whose scholarship application is unsuccessful may still be eligible for a bursary. Scholarships carry additional benefits from the scholarship body, listed in the DfE guidance as membership of the professional body, subject-specialist events, workshops, mentors and coaching, access to classroom resources and professional development.

How the money is paid

The manual says awards for full-time courses are paid in 10 equal monthly instalments over the course. A £20,000 bursary is £2,000 a month and a £22,000 scholarship £2,200 a month. A trainee gets the first payment if actively engaged in the programme on the first day of the month after the start date: a September 2026 starter would be entitled to the first payment on 1 October 2026. Payments are not taxable, though they may be treated as capital income when state benefits are assessed.

Trainees who receive a bursary must not be paid a salary or other payment for teaching work that contributes to the training. They can have both a salary and a bursary if the employment is not related to the course and is not used for assessment against the teachers’ standards. Bursaries are confirmed in writing by the training provider, which the manual says must happen in good time before the course begins.

Other routes and what they pay

An undergraduate bursary of £9,000 is available for an opt-in QTS course in secondary undergraduate languages starting in 2026 to 2027, paid in the final year, or in both the third and fourth years on a four-year course that includes a master’s. A veteran bursary of £40,000 is available to eligible former armed forces members on an eligible QTS course including languages, paid as £20,000 in each of the last two years. Trainees on the School Direct (salaried) route are not eligible for bursaries or scholarships; the DfE instead gives grants to providers, £20,000 for languages including ancient languages in 2026 to 2027.

Teaching English to speakers of other languages abroad uses different credentials, covered in the guide to TEFL and TESOL certification. For how language exams are structured before teaching, see GCSE and A level.

The bottom line

In England, a postgraduate trainee on a QTS course in languages can receive a £20,000 bursary in 2026 to 2027, or a £22,000 scholarship in French, German or Spanish, paid in monthly instalments with the course provider confirming eligibility in writing. Amounts were higher in 2025 to 2026 and may change again. Anyone considering the route should check the current DfE guidance and speak to teacher training providers.

Sources

• Funding: initial teacher training (ITT), academic year 2026 to 2027, Department for Education (gov.uk)
• Initial teacher training bursary funding manual: 2026 to 2027 academic year, Department for Education (gov.uk)
• Funding: initial teacher training (ITT), academic year 2025 to 2026, Department for Education (gov.uk)

Spanish has pulled ahead of French at GCSE in England, while A level entries in the three big languages have been broadly flat for years. The figures come from Ofqual’s “Provisional entries for GCSE, AS and A level: summer 2026 exam series”, supported by Ofqual’s August 2026 blog on GCSE results. They cover England only. The entries are provisional, counting what schools and colleges had submitted to exam boards by 15 April 2026, and final numbers are published later. This article summarises them and is general information.

What the figures count

Ofqual states that the number of entries in each subject corresponds to the number of students entered, except for combined science, other modern languages and ancient languages, where students may enter more than one qualification in the same subject group. In the A level tables, the same applies to modern foreign languages, which Ofqual groups as French, German and Spanish together so that trends can be shown. The release says final entries always vary to some degree from provisional ones.

GCSE entries in summer 2026

Spanish had 132,680 provisional GCSE entries in summer 2026, up 0.5% on 131,985 in 2025. French fell 5.5%, from 128,155 to 121,040. German was almost unchanged at 32,525, up 0.3% from 32,430. Other modern languages fell 3.8% to 41,300, and ancient languages fell 3.5% to 10,960. Ofqual notes that entries for French were among several EBacc subjects that showed a further decline, while Spanish continued to increase. Overall GCSE entries rose 1.1% to 5,840,185.

The five-year series shows how the order changed. In summer 2022, French had 126,185 entries and Spanish 116,355. French was still ahead in 2024, at 130,650 against 129,935. Spanish moved ahead in 2025 and widened the gap in 2026. German has stayed between about 32,000 and 36,000 across the period. Reasons for the shifts are not given in the release, and the DfE evaluation of why pupils opt out of language GCSEs is covered in a separate guide to opting out of language GCSEs.

Higher tier entries and the new GCSEs

French, German and Spanish are tiered, and the release says the percentage of higher tier entries increased for all three compared with the previous year. For French the share rose from 49% to 56%, for German from 59% to 67% and for Spanish from 49% to 55%. In 2022 the higher tier shares had been 53%, 62% and 53%.

Ofqual’s blog adds that summer 2026 was the first year of awards based on the new GCSE courses introduced in 2024, and that overall outcomes in French, German and Spanish were slightly higher in all three subjects than in 2025. It says Ofqual committed to making sure students were neither advantaged nor disadvantaged by being the first to take them. The content of those specifications is explained in the guide to GCSE French, German and Spanish, and the earlier grading adjustment in the guide to grading in French and German.

Other modern languages and early entry

Entries in the “other modern languages” group, which covers GCSEs outside French, German and Spanish, rose from 36,310 in 2022 to 42,945 in 2025 before slipping to 41,300 in 2026. Ofqual notes that in all subjects most entries come from Year 11, but the proportion of entries from students in Year 10 and below is consistently higher in other modern languages than in other subjects. The release does not break the group down by language, so individual languages cannot be separated in these tables.

A level entries

At A level, the provisional 2026 entries were 7,905 for Spanish (up 1.0% from 7,830), 6,465 for French (up 0.2% from 6,450) and 2,190 for German (up 3.1% from 2,125). The combined modern foreign languages figure was 16,555, compared with 16,400 in 2025, 17,035 in 2024, 16,265 in 2023 and 18,750 in 2022. All A level entries rose 2.9% to 845,485 in the same period, so languages remain a small share of the total. Ofqual describes entries in many high-entry subjects as showing slight fluctuations each year.

The A level content that students study is described in the guide to A level modern languages in England.

Reading the numbers with care

Three limits apply. The figures are provisional and rounded to the nearest 5 in Ofqual’s data tables. They count entries submitted by 15 April rather than students who sat each exam. And they describe England only; Wales, Northern Ireland and Scotland have separate qualification systems. The release does not explain the reasons for changes in take-up, so the numbers show what happened, not why.

The bottom line

Ofqual’s provisional figures for England show 132,680 Spanish GCSE entries against 121,040 for French in summer 2026, with German at 32,525. At A level, Spanish, French and German together drew 16,555 entries, close to last year but below 2022. Higher tier shares rose for all three languages in the first year of the new GCSE awards, though the figures are provisional and describe entries rather than final results.

Sources

• Provisional entries for GCSE, AS and A level: summer 2026 exam series, Ofqual (gov.uk)
• GCSE, level 1 and 2 results 2026 at a glance: key trends and context for teachers, Ofqual (ofqual.blog.gov.uk)

Scotland’s school qualifications in languages are built on four skills, each assessed on its own terms, and most of the marks do not come from a single end-of-year paper. This article summarises the National 5 and Higher Modern Languages course specifications, both in their September 2024 editions. Qualifications Scotland replaced the Scottish Qualifications Authority (SQA) on 1 February 2026, and its course pages state that existing documents remain current. The qualifications described here are taken in Scotland; England’s GCSE and A level system works differently. This is general information, and the specifications themselves are the authority.

What the two courses are

Both courses are worth 24 SCQF credit points, with a notional 160 hours of learning including preparation for assessment. National 5 sits at SCQF level 5 and Higher at level 6. The languages offered are Cantonese, French, Gaelic (Learners), German, Italian, Mandarin (Simplified or Traditional), Spanish and Urdu. Each course develops reading, listening, talking and writing in four contexts: society, learning, employability and culture. The Higher specification recommends National 5 Modern Languages, or equivalent qualifications or experience, before starting, though entry is at the centre’s discretion.

National 5: what is assessed

National 5 has five assessment components.

• Reading (30 marks). Candidates read three texts of equal length from different contexts, answer questions in English, and may use a bilingual dictionary. The paper lasts 1 hour 30 minutes.
• Writing (20 marks, scaled to 15). One text of 120 to 150 words responds to a job advert and six bullet points, four of them predictable and two linked to the context. A dictionary is allowed.
• Listening (20 marks, scaled to 30). Candidates hear a monologue worth 8 marks and a conversation worth 12, answer in English, and may not use a dictionary. It lasts about 30 minutes.
• Assignment-writing (20 marks, scaled to 15). A piece of 120 to 200 words on society, learning or culture.
• Performance-talking (30 marks). A presentation worth 10 marks, followed by a conversation on a different context worth 20, of which 5 reward sustaining the conversation.

Higher: a step up in length and complexity

Higher uses the same five components with more demanding tasks. The reading paper has one text, with 20 marks for main points, supporting detail and overall purpose and 10 marks for translating a section into English; candidates have 2 hours for the paper and may use a bilingual dictionary. Directed writing is 150 to 180 words on one of two scenarios, with six bullet points, and assesses appropriate past tenses and at least one other tense, such as the conditional or future. Listening again has a monologue worth 8 marks and a conversation worth 12.

The assignment-writing is 200 to 250 words, mainly discursive, with candidates expected to express or discuss different viewpoints and draw conclusions. The performance-talking is a discussion lasting about 10 minutes on at least two contexts. Together the components total 120 marks. The stage before these courses is covered in the guide to Scotland’s 1+2 language policy covers what schools deliver.

How the coursework is run

In the Higher specification, the assignment-writing is set by centres within the awarding body’s guidelines and written under a high degree of supervision and control in the classroom, with external marking. There is no time limit for its four stages, and candidates may receive feedback, for example through a writing improvement code, before the final piece. The stimulus must stay unseen until the point of assessment, and candidates confirm by signature that the work is their own.

The performance-talking is different. It is conducted and marked in the centre and quality assured externally, and it must be audio or video recorded. The Higher specification calls it a one-off event with only one attempt, though a prelim with different content is allowed. Candidates may use up to five headings of no more than eight words each, as prompts, and the contexts chosen must be communicated beforehand.

Grades and progression

Both specifications grade the course assessment from A to D on the total mark across all components. The Higher specification names progression routes including further study, employment and training, and the Qualifications Scotland subject pages list an Advanced Higher level above Higher. Scotland also offers the Modern Languages for Life and Work awards, which have no exam grade and are covered in a separate guide to Modern Languages for Life and Work. For how European-style levels compare, see CEFR levels explained. The comparable English qualifications are summarised in the guide to GCSE and A level.

The bottom line

Scotland’s National 5 and Higher Modern Languages courses assess reading, listening, writing and talking across four contexts, with a mix of externally set papers, supervised coursework writing and a recorded speaking performance marked in school. Higher lengthens the writing, adds translation into English and expects discursive writing, and both courses grade from A to D. The specifications published by Qualifications Scotland, formerly SQA, are the definitive guide for any candidate or parent.

Sources

• National 5 Modern Languages course specification, version 3.0, September 2024 (sqa.org.uk)
• Higher Modern Languages course specification, version 3.0, September 2024 (sqa.org.uk)
• Advanced Higher Modern Languages course overview, Qualifications Scotland (sqa.org.uk)

How many working-age adults actually speak a second language, and who is most likely to? Eurostat’s “Foreign language skills statistics” article uses the Adult Education Survey (AES) to answer that for the European Union, with data for 2022 and earlier years. This article summarises it. The figures are self-reported, cover adults aged 25 to 64 in EU Member States, do not include the United Kingdom, and are from 2022. They describe people’s own assessment of their skills, not test results, and they are general information.

What counts as a foreign language

Eurostat explains that a language a person knows but does not use at home is treated as foreign. Bilingual people can have more than one native language. In countries with several official languages, such as Belgium with three, a person who speaks only one of them at home counts the others as foreign. People born abroad may treat the language of their country of residence as foreign. Eurostat also warns that countries can apply national preferences when answering the survey, using the example that Slovakian was not counted as foreign in the Czech survey up to 2011.

How many languages adults know

In 2022, 25.3% of working-age adults in the EU said they knew no foreign languages. A further 37.6% knew one, 24.7% knew two and 12.3% knew three or more. Eurostat reports almost no gender gap: 25.5% of men and 25.2% of women knew no foreign language, although a slightly higher share of women knew two (25.5%, 1.6 percentage points above men) or three or more (13.3% against 11.4%).

Country differences were large:

• The share knowing three or more languages peaked in Slovenia (56.3%) and Luxembourg (54.1%), followed by the Netherlands (40.2%) and Finland (38.1%), and was below 4% in Hungary, Greece, Bulgaria and Romania.
• In Romania, Hungary and Bulgaria, more than half of working-age adults knew no foreign language, while ten EU countries had fewer than 10% in that group, led by Slovenia at 4%.
• At least 90% knew a foreign language in ten countries: Slovenia (96.0%), Sweden (95.8%), Estonia (95.5%), Lithuania (95.3%), Latvia (95.0%), Luxembourg (94.5%), the Netherlands (92.9%), Finland (92.1%), Cyprus (91.9%) and Denmark (90.3%).

Eurostat suggests geography and language proximity help explain the high figures in the Nordic, Baltic and some eastern countries, and it says Luxembourg’s figure may partly reflect its three official languages, its education system and the fact that 61.9% of its resident working-age population were born abroad.

Age

Knowing at least one foreign language was more likely for younger, more highly educated and employed adults. In the EU, 84.4% of those aged 25 to 34 knew at least one foreign language, and the share fell in each older group, to 65.0% among those aged 55 to 64. The generation gap was widest where overall knowledge was lower: in Greece, 87.6% of 25 to 34-year-olds knew a foreign language against 51.9% of 55 to 64-year-olds, a 35.7 percentage point gap, with gaps of 29 to 36 points in Poland, Hungary and Romania. Lithuania was the only country with roughly equal shares across the two age groups, and Eurostat suggests that in Lithuania and Latvia, and to a lesser degree Estonia, older people may have learnt Russian while younger people learnt English.

Education

Eurostat reports an education gap too. Across the EU, 90.2% of adults with a high level of education knew at least one foreign language, compared with 71.7% of those with a medium level and 51.2% of those with a low level. Within the medium group, 77.8% of those with a general education knew one, against 73.1% with a vocational orientation. Among people with a low level of education, at least three-quarters knew a foreign language in nine countries, including Denmark, Germany, the Netherlands and Sweden, while fewer than half did in ten, and fewer than 15% in Romania and Hungary.

Employment

By labour status, 78.5% of employed adults knew a foreign language, compared with 65.0% of the unemployed and 63.2% of those outside the labour force. Eurostat notes these figures may support the view that, for some jobs, employers value language skills or promote language learning in staff training, but it presents this as a possibility rather than a finding. In 11 countries, more than 90% of employed adults knew a foreign language, while the share was below 60% in Romania, Bulgaria and Hungary.

Proficiency in the best-known language

The survey also asks about the level reached in a person’s strongest foreign language, and Eurostat stresses that this is self-reported and covers only the strongest language. In 2022, 27.6% of working-age adults who knew at least one foreign language said they were proficient in it, up from 21.6% in 2007. The rise was 1.9, 1.6 and 2.5 percentage points across the three periods in between. In Luxembourg, 68.8% claimed proficiency, and only Latvia (51.5%) and Sweden (51.0%) also passed half. At the other end, Czechia and Italy were both around 15%. Between 2007 and 2022 the proficiency share fell in six of the 22 countries with data, most in Latvia and Slovakia, while Bulgaria, Finland and Cyprus saw relatively rapid increases.

Reading the data carefully

Because answers are self-assessed, they cannot be compared directly with results from a language test, and the article notes that data for some countries may be missing depending on the analysis.

The bottom line

Eurostat’s 2022 figures show that three-quarters of EU working-age adults knew at least one foreign language, with knowledge higher among the young, the highly educated and the employed, and with wide differences between countries. Only about 28% of those who knew a language rated themselves proficient. The findings apply to the EU and are self-reported.

Sources

• Eurostat, “Foreign language skills statistics”, Statistics Explained (ec.europa.eu)

Language courses for adults do not start from nothing: most people in Europe meet at least one foreign language at school. Eurostat’s “Foreign language learning statistics” article (data extracted in April 2026, next update planned for April 2027) reports what pupils in EU countries were learning in 2024. This article summarises the main results. The figures cover EU Member States, not the United Kingdom, and count languages studied at school, not proficiency. They are general information, not a ranking of countries or courses.

What counts as a foreign language

Eurostat notes that some statistics treat official languages as foreign for reporting purposes. In Luxembourg, French and German are counted as foreign languages even though they are official languages, and in Finland students choose between Finnish and Swedish, both counted as foreign. In Belgium, the official languages are Dutch, French and German, which affects how English shares appear. The article covers primary, lower secondary and upper secondary education, and it flags several data notes such as undercoverage in some countries.

Primary school: English first

In 2024, 82.5% of primary pupils in the EU were learning English. Eurostat says learning English is mandatory in primary education in several EU countries, and that pupils tend to receive more instruction in their first foreign language than in later ones, which may magnify English’s apparent importance. Details from the article:

• all or nearly all (99% to 100%) primary pupils in Cyprus, Malta, Austria, Spain and Poland learnt English;
• between 90% and 97% did so in Latvia, France, Greece, Sweden, Finland, Croatia, Italy and Romania;
• in Belgium the share was 23.0%, reflecting the focus on other official state languages; and
• Luxembourg generally did not teach English in primary school, focusing on German, French and Luxembourgish.

Only 4.4% of EU primary pupils were learning French and 4.0% German. Luxembourg reported the highest shares for both (79.9% and 94.7%). Among primary pupils, 6.4% were learning two or more foreign languages, and Luxembourg (78.7%) was the only country where a majority were. The next highest were Latvia (41.4%), Greece (35.5%), Denmark (29.0%) and Finland (28.6%). Eurostat records that the EU share of primary pupils learning at least two languages rose from 4.6% in 2014 to 6.4% in 2024. It also notes that no foreign languages were taught in Irish primary education before September 2025, when a phased introduction of modern foreign languages began.

Lower secondary: more languages, more variety

By lower secondary level, 97.6% of EU pupils were learning English, virtually unchanged from 97.7% in 2014. All lower secondary pupils in Denmark, Slovenia and Sweden were learning English. Other languages were far more varied. In 2024, 30.3% of EU lower secondary pupils were learning French, 19.2% Spanish and 18.9% German.

Eurostat highlights some national patterns:

• French: Luxembourg had more than four fifths (82.6%), and Romania, Cyprus, Portugal, Italy, the Netherlands, Belgium and Greece all had more than half;
• Spanish: France was the only country with a majority (56.6%), followed by Sweden (47.0%); and
• German: Denmark (74.9%) and Luxembourg (71.9%) had large shares, with no other country above half.

Since 2014, Spanish gained (from 13.1% to 19.2%), while German fell by 4.1 percentage points and French by 3.4. Overall, 58.8% of lower secondary pupils learnt two or more foreign languages in 2024, down from 59.7% in 2014. Finland had the highest share (97.2%), while Ireland (5.2%), Hungary (5.6%) and Austria (7.5%) had the lowest.

Upper secondary: general versus vocational

Eurostat’s headline is that 89% of upper secondary pupils in the EU learnt English in 2024, including 96% in general programmes and 81% in vocational programmes. In Malta and Sweden all pupils learnt it. Denmark (50.8%) was the only country below 70%. Nearly half (49%) learnt two or more foreign languages.

Beyond English, slightly fewer than a fifth of upper secondary pupils learnt German (19.3%), Spanish (17.9%) or French (17.2%). Majorities were learning German in Luxembourg (69.6%) and Poland (62.1%), Spanish in France (60.6%), and French in Romania (84.7%), Luxembourg (70.5%), Cyprus (69.2%) and Belgium (52.6%). The gap between programme types was largest for Spanish: 27.6% of pupils in general programmes against 7.0% in vocational programmes.

For English in upper secondary education, the largest gains between 2014 and 2024 were in Germany (up 11.6 percentage points), Belgium (up 11.5) and Greece (up 10.6). Eurostat also notes that Portugal was the only country where the share learning English was higher in vocational than in general programmes (by 7.9 percentage points). For German, the EU share was 20.2% in general programmes and 18.0% in vocational ones, and Poland and Italy were the only countries where vocational shares were higher. For French, the shares were 20.1% in general and 14.7% in vocational programmes, and in Greece 32.9% of general-programme pupils learnt it against 0.7% in vocational programmes.

Interpreting the figures

Eurostat’s counts show who studies a language, not how well they learn it, and the article does not measure outcomes. Its data notes also warn that some figures are affected by undercoverage or by ancient languages being included, so small differences between countries should not be over-read.

The bottom line

Eurostat’s 2024 figures show English as the near-universal first foreign language in EU schools, with French, German and Spanish taught to much smaller shares and in strongly national patterns, and with a large minority of secondary pupils studying two or more languages. They describe EU school systems only, not adult learning and not the UK.

Sources

• Eurostat, “Foreign language learning statistics”, Statistics Explained (ec.europa.eu)

Should a language class be conducted entirely in the language being taught? Are real newspaper articles better than adapted ones? Is it better to be corrected on every mistake? Ofsted’s 2021 research review on languages tackles these perennial questions in its section on pedagogy, drawing on second language acquisition research. This article summarises that section and Ofsted’s announcement of the review. The review concerns school language teaching in England, describes research and debate rather than rules, and says no single approach achieves high-quality provision. Adult learners can read it as background to how classes are designed, not as a guarantee about any particular course. It is general information.

Intentional and incidental learning

The review distinguishes two kinds of learning. Intentional learning happens when the learner is aware of the need to learn, invests effort and knowingly gains explicit knowledge. Incidental learning happens as a by-product of another activity, without the learner realising. Researchers debate whether languages are learned explicitly or implicitly, and the review says the general consensus is that some of both goes on in most situations.

Incidental learning depends on individual differences such as how analytical a learner is, how engaged with the task, and how good their working memory is. The review says that when time is short and exposure to the language is limited, learners are less likely to meet the same phonics, vocabulary and grammar often enough to notice and remember them incidentally. It concludes that, for most learners in most situations, it is likely to be efficient for teachers to promote intentional learning, and that there is little evidence in second-language acquisition research that words or structures learned incidentally are retained better than those learned intentionally. It links this to the expertise reversal effect: explicit teaching tends to work best with novices, while enquiry-based approaches can succeed with advanced learners.

How much target language should the teacher use?

The review calls the role of the target language a perennial debate. It says learners should be exposed to the language but not overwhelmed in the early stages, to the point of demotivation. It also says the teacher’s use of the language should not stop pupils from understanding how the language is structured, while accepting that using it is essential for practice and reinforcement, including familiarity with rhythm, sounds and intonation.

Its conclusions are practical:

• the teacher’s use of the target language should be carefully planned within the scheme of work and build systematically on prior knowledge, with English used when needed;
• the earlier the stage, the more care is needed to avoid overly complex language, and a concrete example is using the target language for classroom routines that learners clearly understand, rather than for explaining complex grammar; and
• learners’ own use of the target language should be central, since the review says there is some evidence linking teachers’ widespread use with effective learning, but clear evidence that pupils’ own use of it positively affects learning.

For learners to say what they actually mean, not just repeat formulaic routines, the review says they need the linguistic capability and the motivation for real speech, and the classroom should give them chances to try the language out.

Are authentic texts better?

An authentic text is one taken largely unamended from a native-speaker environment, such as an article from a French newspaper or website. The review says it is not well evidenced that authenticity alone helps learning or motivation. Texts should support the sequence of vocabulary and grammar being taught, and adapted authentic texts or bespoke ones can serve the purpose. Authentic material can still offer chances for pupils to follow their interests, handle new materials and learn about culture.

The review cites research suggesting that learners need to know around 95% of the words in a written or audio text for adequate comprehension, so teachers must choose resources of suitable difficulty. For listening, it points to the real-time nature of speech, and suggests that presenting audio and written text together can help learners split the sound stream into words, understand and learn vocabulary.

Correcting errors

The review says feedback works best when teachers give salient, focused and clear feedback and pupils are willing and able to notice and use it. It acknowledges a concern that correction can demotivate learners from attempting more complex language, while noting broad consensus that focused correction, clearly understood by pupils, can benefit many. It describes three strategies:

• recasting, where the teacher restates what the pupil said with the error corrected;
• prompting, where the correction is elicited from the learner; and
• explanation, where explicit, often technical, information is given about the rule behind the error.

According to the review, prompting forces the learner to retrieve the language and seems more likely than recasting to be effective for most learners, most of the time, for most types of error, with pronunciation a possible exception that may be best corrected by a direct recast. Recasting relies on learners working out what the error was, and even when they repeat the correction it is not clear that this leads to long-term learning. It adds that early-stage learners may benefit from rule-based correction.

On timing and scope, the review cites a view that correction is most effective when immediate, and evidence that focused correction of one or a few elements is usually more effective than correcting every error. It also says classrooms should treat mistakes as a normal, helpful part of learning.

The bottom line

Ofsted’s review favours planned, intentional teaching, a carefully calibrated amount of teacher target language with plenty of learner speech, texts chosen for what they teach rather than for authenticity alone, and focused correction that prompts learners to retrieve the right form. It is a literature review for schools in England and leaves room for different approaches.

Sources

• GOV.UK, Ofsted, “Research review series: languages”
• GOV.UK, Ofsted, “Ofsted publishes research review on languages education”

Any language course, whether a school scheme of work or an evening class, has to decide what to teach first. Ofsted’s research review on languages, published on 7 June 2021, sets out the evidence it draws on for that decision in England’s schools. This article summarises the review’s account of curriculum progression: the “three pillars” of phonics, vocabulary and grammar. It relies on the research review itself and on Ofsted’s announcement of it. The review applies to England, describes a review of the literature rather than inspection rules, and says there is no single way of achieving high-quality languages education. It is general information for learners and parents, not a recommendation of any course.

Why Ofsted reviewed the evidence

Ofsted’s announcement says the review explores the literature on foreign languages education to identify what contributes to a high-quality languages curriculum, assessment and teaching. It says good curriculum design is crucial to ensuring that pupils securely learn the vocabulary, grammar and phonics that will let them manipulate the language for themselves. The review’s national context notes that the number of pupils taking a language at GCSE fell significantly after 2004, when studying a language after age 14 stopped being compulsory, and that a language has been required in primary schools from age 7 since 2014.

The three pillars

The review says a language curriculum must be planned by considering the building blocks of the subject and their sequence. For languages, those blocks are the sounds, the words and the rules connecting them into sentences and meanings. It calls them phonics, vocabulary and grammar, and says assessment systems typically incorporate all three. It stresses that this is not a reductive approach: the goals of conversing fluently and exploring cultures can only be reached on firm foundations. It also says language is more than the sum of its parts, and that learners need other competencies to communicate with different people in different contexts.

Phonics

The review says clear pronunciation and the link between sounds and spelling are integral to second language learning. Strong phonological awareness, such as hearing the difference between “back” and “pack”, supports listening, and being able to decode written words into sounds helps reading, autonomy and vocabulary learning. It also says there is little strong evidence for one prescribed order for teaching sound-spelling relations in fine detail, but that teaching needs to draw explicit attention to them. Knowledge of a learner’s first-language sound-spelling system can strongly influence learning a second. The review notes there are significantly more unfamiliar sound-spelling correspondences in French than in Spanish or German for English speakers.

Vocabulary

Vocabulary is described as crucial: studies show a wider vocabulary correlates with other aspects of language ability, including reading and grammatical awareness. The review says vocabulary choice should be considered carefully with regard to age and word frequency. It cites several studies estimating that the 2,000 most common words of a language make up more than 80% of the words in most written and spoken texts, and that high-frequency words act as anchor points for navigating texts.

Less frequent words can still be useful for an individual learner’s needs. Many textbooks organise words by topic, such as hobbies or holidays, and the review accepts that this is practical but says learners need to use words across many contexts. It gives three reasons: meaning and use vary with context, words are easier to remember if met in several topics, and exam boards expect learners to transfer topic vocabulary to other themes.

The review also divides word knowledge into three components:

• breadth, meaning the number of words known;
• depth, meaning whether pupils understand a word and can use it appropriately, know its forms across gender and number, and know related synonyms and antonyms; and
• fluency, meaning the speed and automaticity of recall.

Curriculum plans should let learners use new words in comprehension and production, in both spoken and written forms. The review says practising across modalities is likely to embed vocabulary better and increase fluency.

Grammar

Grammatical progression, the review says, needs careful planning over time so that pupils do not become confused or demotivated as structures grow more complex. It expects a logic running from simpler to more complex concepts and lists decisions about when to introduce:

• different tenses;
• verb agreement for person and number;
• agreement on nouns and adjectives;
• negation;
• question forms; and
• different parts of a verb paradigm, since giving a whole paradigm early may overwhelm working memory.

The review adds that GCSE and A-level specifications contain a substantial amount of grammar and that the Common European Framework of Reference (CEFR) outlines grammatical progression across ability levels.

Learning principles behind the sequencing

The review’s guiding principles come from cognitive science and Ofsted’s inspection framework research. Among them, it says short-term memory has limited capacity, so overloading it with too much at once limits learning. Spaced or distributed practice, where knowledge is rehearsed briefly over a longer period, is described as more effective than massed practice, and retrieval practice, recalling something previously learned, helps retention. It also mentions the expertise reversal effect: explicit teaching works best for novices, while enquiry-based approaches can work for experts.

What this means for choosing a course

The review is aimed at schools, so its features of high-quality education are not a checklist for adult classes. Even so, they suggest questions any learner can ask of a course: whether pronunciation and spelling are taught explicitly, whether the most frequent words are covered early, whether vocabulary returns in new contexts, and whether grammar is introduced in a planned order with regular review.

The bottom line

Ofsted’s 2021 review treats phonics, vocabulary and grammar as the pillars of language progression, with high-frequency words, deep and fluent word knowledge and carefully staged grammar as central. It rests on a review of research, not on a single trial, and it applies to England’s school curriculum.

Sources

• GOV.UK, Ofsted, “Research review series: languages”
• GOV.UK, Ofsted, “Ofsted publishes research review on languages education”

Anyone choosing a language course, whether to learn a new language or to strengthen one spoken at home, is entering a country where many languages are already in daily use. The Office for National Statistics (ONS) published the first detailed picture from Census 2021 on 29 November 2022 in “Language, England and Wales: Census 2021”, and defines the term it uses in its Census 2021 dictionary. This article summarises those two pages. The figures cover England and Wales only and describe 21 March 2021, so they are a snapshot rather than a current count. Scotland’s census was held on 20 March 2022 and Northern Ireland’s on 21 March 2021, and neither is covered here. It is general information.

What “main language” means

ONS defines main language as a person’s first or preferred language. The census question covered usual residents aged three and over. In Wales, respondents were asked whether their main language was anything other than English or Welsh, so ONS says it is not possible to work out how many people in Wales consider Welsh to be their main language. Welsh language skills were the subject of a separate question in Wales, and the Welsh Government was to publish that information on 6 December 2022. This distinction matters when comparing the headline figures with Welsh-language learning.

The headline figures

In England and Wales, 91.1% of usual residents aged three and over (52.6 million of 57.7 million) had English, or Welsh in Wales, as a main language. That was down from 92.3% (49.8 million) in 2011. A further 7.1% (4.1 million) were proficient in English but did not have it as their main language.

The most common main languages other than English, or Welsh in Wales, were:

• Polish, 1.1% (612,000), which remained the most common;
• Romanian, 0.8% (472,000), which moved into the top 10 and recorded the largest increase over the decade, from 0.1% (68,000) in 2011;
• Panjabi, 0.5% (291,000); and
• Urdu, 0.5% (270,000).

ONS says the rise in Romanian mirrors increases between the two censuses in the number of people born in Romania and in those choosing only a Romanian national identity.

Differences across the two nations and regions

English, or Welsh in Wales, was the main language of 96.7% of people in Wales (2.9 million of 3.0 million) and of 90.8% in England (49.7 million of 54.7 million). Within England, the North East had the highest proportion with English as a main language (96.5%, 2.5 million), while London had the lowest (78.4%, 6.7 million).

Language patterns were local as well as regional:

• Polish: the highest regional share was the East Midlands (1.5%, 71,000). Across local authorities it ranged from 0.1% (100) in Castle Point to 5.7% (4,000) in Boston.
• Romanian: London had the highest share (1.9%, 159,000), and Harrow the highest local share (7.5%, 19,000).
• Panjabi: the West Midlands had the highest regional share (1.4%, 83,000), and Wolverhampton the highest local share (6.5%, 17,000).
• Urdu: the North West had the highest regional share (0.8%, 59,000), and Slough the highest local share (4.3%, 7,000).

In Wales, Polish was again the second most common main language after English or Welsh (0.7%, 21,000), with the highest local share in Wrexham (2.5%, 3,000). Arabic was next (0.3%, 9,000), highest in Cardiff (1.4%, 5,000).

British Sign Language

The census also recorded British Sign Language (BSL) as the main language of 22,000 people (0.04%) aged three and over across England and Wales, up from 15,000 in 2011. ONS notes that Derby had the highest share, 0.2% (400 people), and suggests that the presence of the Royal School for the Deaf could be one reason. Within England the figure was just over 21,000 and in Wales just over 900.

English proficiency

Among the 5.1 million people (8.9% of usual residents) who did not have English, or Welsh in Wales, as a main language, ONS reports that 43.9% (2.3 million) spoke English very well and 35.8% (1.8 million) spoke it well. About 17.1% (880,000) could not speak English well and 3.1% (161,000) could not speak it at all. Expressed against the whole population, 1.5% could not speak English well and 0.3% (161,000) could not speak it at all. For households without English as a main language, the census offered interpretation services and translated leaflets in more than 50 languages.

Households

The census asked, for the first time, whether household members shared a main language. Of 24.8 million households, 69.8% had more than one person. Among those, all members shared the same main language in 15.8 million households (63.8%). Languages differed between generations but not within partnerships in 480,000 households (1.9%), differed within partnerships in 534,000 (2.2%), and differed between other relationships, such as friends, in 465,000 (1.9%).

Reading the numbers carefully

ONS says the person response rate was 97% of the usual resident population, with over 88% in all local authorities.

The bottom line

Census 2021 shows that, in England and Wales, about nine in ten residents had English or Welsh as a main language, with Polish, Romanian, Panjabi and Urdu the most common others and strong local concentrations in places such as Boston, Harrow, Wolverhampton and Slough. Anyone weighing a language course, or a heritage-language class, can use those patterns as background, remembering the figures are from 2021 and only for England and Wales.

Sources

• Office for National Statistics, “Language, England and Wales: Census 2021” (ons.gov.uk)
• Office for National Statistics, “Main language variable: Census 2021” (ons.gov.uk)