Executive verdict
Dutch people switch to English for several reasons: an attempt to help, a need for faster communication, an established relationship habit, or an assumption about which language everyone understands. Learner accounts document these explanations, but there is no verified national percentage showing how often switching happens or which motive dominates.
The English proficiency behind the phenomenon is substantial. The Netherlands scores 624 and ranks first among 123 countries and regions in EF EPI 2025 [1][2]. However, EF explicitly describes its test-taking population as self-selected and not necessarily representative [3]. A proficiency ranking cannot tell you why a particular cashier, colleague or neighbour changed languages.
The frustration is also documented beyond internet discussion. Utrecht's municipal research records a learner finding an unsolicited switch discouraging, while Leiden's student platform describes English switching as a barrier to practising Dutch [4][5]. These are accounts of experience, not national prevalence estimates.
Our verdict: you cannot reliably prevent every switch, but you can make your preference clearer and arrange more dependable opportunities to speak. Ask to continue in Dutch, explain what help you need, and choose situations where the other person has time. Treat a switch as information about the conversation, rather than an assessment of your intelligence, effort or language level.
How good is English proficiency in the Netherlands, really?
English proficiency is high by international comparison, but there is no single percentage that means everyone is fluent. Test scores, self-reported conversational ability and everyday willingness to use English measure different things.
EF's country page and downloadable Netherlands factsheet agree on the headline score of 624 [1][6]. Its international ranking covers 123 countries and regions, using results from approximately 2.2 million adults [2]. That is extensive coverage, but it is not a census.
The following figures answer different questions and should retain their original units.
| Indicator | Published result | What it measures |
|---|---|---|
| Netherlands international position | First among 123 countries and regions in EF EPI 2025 [1][2] | Relative performance within EF's participating countries and regions |
| Netherlands overall EF score | 624 in the 2025 edition [1] | A country index score, not the percentage speaking English |
| EF reading and speaking scores | Reading 634; speaking 514 [6] | Separate skill results, not a guarantee of balanced individual proficiency |
| Eurobarometer data reanalysis | 85.9% classified as English speakers using data published in 2024 [7] | Native speakers plus foreign-language speakers reporting good or very good competence; basic competence is excluded [8] |
Why the definitions matter
The independent language knowledge /eu website explains that its threshold excludes respondents describing their foreign-language competence as basic. Consequently, its results differ from broader Eurobarometer conversational-ability figures that include basic speakers. The underlying survey took place in autumn 2023 [8]. Its 85.9% figure is the site's published reanalysis, not a new Dutch government fluency estimate [7][8].
EF also requires caution. Its methodology uses a rolling calculation combining the previous calendar year's test average with the preceding published country scores. The sample favours younger adults interested in language assessment [3]. Calling the result a fresh, representative survey of all Dutch residents would misdescribe it.
Our calculation from the Netherlands factsheet gives a reading-to-speaking difference of 120 index points, subtracting 514 from 634 [6]. This is a difference between published aggregate skill scores. It does not mean that every participant has that gap, or that a Dutch speaker's English is automatically better than your Dutch.
Marinel Gerritsen's review of English in Dutch-speaking Europe likewise distinguishes the strong overall position of English from variation between people and the limitations of self-assessment [9].
For learners, the useful conclusion is modest: English is sufficiently available to become an easy alternative. Availability does not establish anyone's motive, patience or preferred language.
Why do Dutch people switch to English?
Switching can solve an immediate communication problem while creating a practice problem for the learner. The evidence supports several plausible explanations, rather than a single national personality trait.
They think they are helping
In a learner discussion, contributors describe switching because someone appears to struggle or because they want to avoid excluding them. Others report that explaining a preference for Dutch changes the interaction [10]. These are self-selected accounts, but they identify a practical misunderstanding: the speaker assumes that easier communication is the help being requested.
Your actual goal may be different. You might accept slower communication because speaking Dutch is part of the purpose of the encounter. Unless you say so, that goal can remain invisible.
This suggests a useful distinction between helping someone finish a task and helping someone practise. A person can sincerely intend the former while accidentally obstructing the latter.
The situation rewards speed or inclusion
Leiden's student report describes both limited Dutch practice because people switch to English and difficulties joining Dutch-speaking social settings [5]. The apparent contradiction disappears when the setting changes: a brief exchange can accommodate an international easily, while an established group may continue using its shared language.
Our inference is that context matters more than slogans about Dutch hospitality. A service interaction, a social gathering and an agreed practice session have different purposes. Your request is more workable when it fits the purpose of the encounter.
English has become the relationship's default
An established language can continue even after the learner's ability changes. Contributors to the learner discussion describe automatic English use with people they first met in an international setting [10].
There is institutional testimony to this habit beyond the Netherlands. In an interview published on 15 October 2026, Huis van het Nederlands Brussel director Patrick Manghelinckx describes switching despite knowing that someone wants to speak Dutch [11]. This is an experienced practitioner's observation, not an experiment establishing how common the behaviour is.
The practical implication is to negotiate a replacement routine. A vague intention to speak more Dutch leaves the previous habit intact.
Assumptions can arrive before your Dutch does
The same learner discussion includes accounts of being addressed in English before speaking [10]. Such reports cannot establish the prevalence or cause of appearance-based assumptions. They do, however, challenge the idea that every English response must have been triggered by a mistake in the learner's Dutch.
You do not need to diagnose a stranger's motive before stating your preference.
Does switching mean my Dutch or accent is bad?
No. A switch is an interactional choice, not a reliable language assessment. It can accompany a genuine misunderstanding, but it can also occur because of habit, uncertainty or assumptions.
Pronunciation research provides an important distinction. A meta-analysis published in 2025 combined 49 studies and found a relatively weak relationship between accentedness and intelligibility compared with several other speech dimensions [12]. The authors support prioritising understandable speech rather than treating accent reduction as the overriding goal.
An earlier retrospective analysis by Murray Munro and Tracey Derwing also emphasises that heavily accented speech can remain highly intelligible [13]. These findings concern second-language speech generally. They do not quantify how a particular Dutch accent affects the probability of an English switch.
Diagnose the problem you can actually improve
For practical purposes, separate these questions:
- Did the listener identify the words you intended?
- Did you understand the reply?
- Could you continue when something unexpected happened?
- Had you agreed which language to use?
These are our proposed diagnostic questions, not a validated switching test. Their value is that they lead to different actions.
If a word repeatedly fails to land, work on its pronunciation. If the reply causes difficulty, practise listening and asking for clarification. If communication succeeds but English keeps returning, revisit the language agreement.
A useful tutor instruction is to identify where meaning breaks down and practise repairing that moment. General advice to sound more Dutch is less specific and harder to evaluate.
There is no level at which switching must stop
The CEFR describes developing capabilities, not other people's language choices. Its A2 interaction descriptor covers routine exchanges while allowing difficulty sustaining a conversation. B1 includes unprepared conversation on familiar topics, while B2 describes greater fluency and spontaneity [14].
Those descriptors support matching tasks to your current ability. They do not establish a threshold below which you should avoid speaking or above which everyone will continue in Dutch.
A rehearsed opening may be within reach before an unpredictable follow-up. Prepare both the opening and a way to request help. Successful participation can include clarification, hesitation and reformulation.
What do learner reports actually show?
Learner frustration is documented, but the available reports do not establish a national switching rate or prove that switching causes a particular learning delay. Their strongest contribution is showing how English availability interacts with confidence and access to conversation.
Utrecht's integration research received 296 questionnaire responses, an 18% response rate, and included interviews with 10 integration participants. The municipality explicitly warns that the respondents are not representative of all local integration participants [15].
Its language chapter reports the following results for the surveyed groups [4].
| Experience reported in Utrecht's 2025 report | Asylum status holders | Family migrants |
|---|---|---|
| Find speaking Dutch daunting | 76% [4] | 70% [4] |
| Experience everyday problems because speaking Dutch is difficult | 63% [4] | 49% [4] |
| Are satisfied or very satisfied with their spoken Dutch level | 38% [4] | 27% [4] |
These figures concern speaking experiences. None is the share whose conversation partners switch to English.
The municipal summary independently reproduces the satisfaction figures and describes learners actively seeking opportunities to practise [16]. The language chapter also records English dominating study, work and home life, alongside an account of frustration when a Dutch conversation becomes English [4].
Our inference is that an English-speaking environment can reduce the practical necessity to use Dutch while leaving the desire to participate in Dutch-speaking life unresolved. That is a plausible mechanism, not a measured causal effect.
Leiden's student platform reports a similar tension, while warning that its student responses cannot represent the whole student population [5]. Its findings should be read as local testimony and proposals for improvement.
These limitations do not make the experiences unimportant. They determine which claim is justified: the barrier exists for some learners. The evidence does not tell us how frequently every international encounters it, or how much faster they would learn if it disappeared.
How do I stop Dutch people switching to English?
Ask explicitly to continue in Dutch, then make it easier to maintain that choice. The wording below is our suggested practice script, not a set of experimentally proven formulas.
State the preference before a misunderstanding develops
A useful opening is:
Ik leer Nederlands en wil graag oefenen. Kunnen we Nederlands blijven spreken? (I am learning Dutch and would like to practise. Can we keep speaking Dutch?)
If someone has already changed languages:
Dank je, maar ik wil het graag in het Nederlands proberen. (Thank you, but I would like to try it in Dutch.)
The learner discussion contains accounts of explicit requests helping, alongside explanations of why speakers had switched [10]. That supports trying the request. It does not justify promising that everyone will agree.
A preference is clearer than an apology. You do not have to begin by declaring that your Dutch is terrible, explaining your entire learning history, or comparing your ability with the other person's English.
Ask for the particular help you need
Our suggested alternatives are:
- Kun je dat iets langzamer zeggen? (Can you say that a little more slowly?)
- Kun je dat met andere woorden zeggen? (Can you say that in different words?)
- Wat betekent dat woord? (What does that word mean?)
- Ik begrijp je, maar ik heb even tijd nodig om te antwoorden. (I understand you, but I need a moment to answer.)
The last sentence is particularly useful when comprehension and speaking speed differ. It tells your partner that a pause does not necessarily mean their Dutch reply was incomprehensible.
For more formal encounters, use u (the formal form of you) rather than je (the informal form of you), adapting the verb accordingly.
Keep an agreed route back to Dutch
If a missing word requires English, propose using it briefly and returning to Dutch. Our suggested wording is:
Hoe zeg je dat in het Nederlands? Daarna probeer ik het opnieuw. (How do you say that in Dutch? Then I will try again.)
The aim is continuity. A useful explanation in English need not decide the language of the rest of the encounter.
Should you pretend you cannot speak English?
A learner thread proposes exactly this tactic and contains both supportive responses and suggestions to ask directly instead [17]. It is evidence that learners discuss the strategy, not evidence of its comparative effectiveness.
Our recommendation is an honest request. Pretending can create an unnecessary problem if you subsequently need English, and it leaves your actual learning preference unexplained.
Consider an agreed bilingual conversation
Luistertaal (receptive multilingualism) means participants speak different languages while understanding each other. Utrecht University research commissioned by Taalunie documents its use and recommends explicit language agreements and checks on mutual understanding [18].
For a learner, this offers flexible options: speak Dutch while accepting English replies to practise production, or listen to Dutch while answering in English to practise comprehension. Explain the arrangement first. The choice should support a shared conversation rather than become a contest over whose language prevails.
Where can I practise Dutch, including in Amsterdam?
Recurring conversations with willing partners provide more dependable practice than relying entirely on spontaneous encounters. Amsterdam is not automatically unsuitable, and moving to a smaller town is not an evidence-based guarantee.
EF's city results place Groningen at 656 and Amsterdam at 630 [1]. These are English proficiency scores, not measurements of how readily residents accommodate Dutch learners. They cannot establish which city will give you more Dutch conversation.
A better selection criterion is whether speaking Dutch is an understood part of the activity.
Use settings where practice is already expected
Het Begint met Taal directs learners to local conversation cafés, language coaches, online practice and workplace support. It also makes clear that conversation coaches are not language teachers [19]. A taalcafé (language conversation café) or taalmaatje (language buddy) makes the purpose visible before the conversation starts.
Kletsmaatjes describes an online format based on an hour each week, with volunteers supporting listening and speaking [20]. This is a published programme format, not a claim that an hour is sufficient for a particular learning outcome.
WorkinNL also identifies libraries, a Taalhuis (local language support centre) and volunteer language coaching as routes to practice, including free support [21]. Check local availability and eligibility rather than assuming every advertised activity accepts every learner immediately.
Make ordinary encounters manageable
Our recommendation is to choose familiar activities with enough room for conversation: a regular hobby, a neighbour you already know, or a quieter moment with a familiar service provider.
Prepare the task's vocabulary and likely follow-up questions. Ordering successfully is useful, but asking about an alternative or explaining a preference creates a different speaking challenge.
If a staff member is busy, complete the transaction efficiently and look elsewhere for extended practice. If they do not speak Dutch, another setting is needed.
Use the directory of Dutch schools, courses and resources to find structured options, or compare online Dutch tutors when you need scheduled speaking time and teaching feedback.
How can I practise Dutch with colleagues or my partner?
Agree on a specific activity in Dutch and decide what to do when communication becomes difficult. An established relationship needs a workable routine more than a repeated general request to speak more Dutch.
At work, distinguish practice from participation
Study in NL, Nuffic's official information platform, advises learning Dutch early and explains that Dutch remains the country's main language and can be required for employment [22]. That makes workplace practice valuable, even where the current team operates in English.
WorkinNL identifies workplace language buddies as an option and notes that some employers offer language lessons, sometimes during working hours [21]. Availability depends on the employer.
Our proposed arrangement is to select a recurring, manageable activity: explaining a completed task, discussing a familiar document, or having a willing colleague continue a casual conversation in Dutch.
Agree how corrections work. Immediate correction may help during deliberate practice, while a note afterwards may preserve the flow of ordinary conversation. Ask the learner which they want.
For meetings, consider everyone present. A learning preference should be accommodated without making another colleague unable to participate. Taalunie's guidance on receptive multilingualism recommends checking understanding and agreeing alternatives when the chosen arrangement fails [18].
For task-specific preparation, the business Dutch guide can help you define what you actually need to do in the language.
At home, keep the relationship in view
Our recommendation for couples is to attach Dutch to a familiar shared activity, such as cooking, walking or discussing plans. Keep an agreed way to pause practice when the conversation becomes emotionally or practically demanding.
A partner can be a willing conversation partner without becoming your teacher. Decide whether you want correction, help finding words, or simply time to finish.
When extended family automatically uses English, a partner can explain your preference before the gathering. That shares the effort of negotiating language instead of leaving you to restart the discussion with every person.
Is the English trap different in Flanders and Brussels?
English availability is also substantial in Dutch-speaking Belgium, but evidence from Belgium, Flanders and Brussels must remain geographically distinct. None provides a shortcut to predicting an individual's language choice.
EF reports Belgium's overall score as 608 and its Dutch-speaking regional score as 654 [23]. The national Belgian result therefore conceals substantial regional variation. It is not a direct estimate for Flanders, and neither score measures switching.
Huis van het Nederlands Brussel's motivation research draws on surveys of 4,139 unique learners, alongside focus groups and interviews. The surveys took place in late 2020 or early 2021 and summer 2022 [24].
The report describes disappointment about limited opportunities to use Dutch and develop Dutch-speaking contacts, including among advanced learners [24]. This is a wider participation issue than the language chosen during a brief transaction.
Manghelinckx's later interview distinguishes organised practice opportunities from opportunities to use Dutch in ordinary life [11]. That distinction also helps interpret the Netherlands experience.
Our inference is that learners benefit from both. A conversation café makes speaking accessible, while a recurring role in a club, workplace or neighbourhood gives Dutch a purpose beyond practice. Brussels evidence supports examining that distinction, but should not be treated as a measurement of Dutch learners' experiences in Amsterdam.
What this means for learners
Build a routine that includes an explicit request, manageable conversation tasks and dependable practice. Judge progress by what you can communicate and understand, rather than by whether strangers ever use English.
The following is our practical framework, derived from the evidence but not tested as a complete intervention.
Make your preference easy to understand. Choose a Dutch sentence asking to continue in Dutch and practise saying it comfortably. Include what you need, such as slower speech or time to answer.
Choose a real task. Prepare vocabulary for something you actually do. Practise the likely reply as well as your opening. Afterwards, identify the unexpected part that prevented you continuing.
Arrange recurring contact. Find a willing buddy, conversation group, colleague or teacher. Specify that you want speaking time. An English-friendly environment is easier to navigate when your practice does not depend entirely on chance.
Keep a brief record of outcomes. Note whether you completed the task, understood the reply, requested clarification and continued after a difficulty. These observations are more actionable than recording only that someone switched.
Use feedback selectively. Work on errors that repeatedly obstruct meaning. Keep broader grammar and pronunciation development in your lessons without demanding perfection before participating.
Ask established speakers to share responsibility. If you are the Dutch-speaking colleague or partner, ask what language the learner prefers. Allow pauses, clarify meaning and avoid turning every contribution into a correction exercise.
A learner can prepare and persist, but opportunities are co-produced. Hosts, colleagues and institutions also determine whether Dutch is available for meaningful participation. That is the practical implication of the local reports and the guidance on agreed language use [5][18][24].
What we could not verify
A national switching rate. We found no representative Dutch study establishing the proportion of encounters that move from learner Dutch to English. EF proficiency results cannot fill that gap.
A single dominant motive. Learner discussions and practitioner accounts identify plausible explanations, but cannot establish their national distribution. An unsolicited switch cannot reliably reveal whether the speaker intended help, efficiency, exclusion or something else.
A guaranteed phrase or proficiency threshold. We found no controlled comparison showing that a particular request consistently prevents switching. CEFR descriptors specify communicative capabilities, not other people's responses [14].
A reliable city ranking for practice. We could verify EF city proficiency scores, but not comparable city-level data on opportunities for learner Dutch [1]. Advice that smaller towns always solve the problem remains unverified.
The learning time lost to English. The available accounts do not establish how much switching delays Dutch acquisition. Brief English support may also allow a conversation to continue.
A fresh official population fluency percentage. We report the accessible Eurobarometer reanalysis with its stated threshold. We did not independently reproduce its weighted calculations from respondent-level data [7][8].
The recommended scripts and routines are therefore informed practical proposals. They should be adapted to the people involved and evaluated by whether they create useful, mutually workable conversation.