The evidence question
Is Your English Worse When Someone Is Listening?
Why can your English feel fluent alone and suddenly disappear in a real conversation? Research on foreign-language anxiety, working memory, speaking performance, and accent concerns helps explain the gap.
A question worth asking.
08 sources, open to exploreIn this article 18 sections
You know the word.
You have used it before.
You can probably say the whole sentence perfectly in your head.
Then another person looks at you and waits for your answer.
And suddenly:
I… uh… how do you say…
The word is gone.
Five minutes later, when the conversation is over, it comes back immediately.
This experience can make language learners think something strange happened to their English.
Maybe they are not as fluent as they thought.
Maybe their vocabulary is weaker than it seemed.
Maybe they need to study more grammar.
Sometimes those explanations are true.
But there is another possibility:
the English you know and the English you can access under pressure are not always the same thing.
Research on foreign-language anxiety suggests that being observed, evaluated, or put under communicative pressure can temporarily change how efficiently you process and produce a second language.
That does not mean anxiety magically deletes vocabulary.
It means speaking is already a demanding cognitive task — and anxiety can compete for some of the same mental resources.
Foreign-language anxiety is not just ordinary nervousness
Researchers have studied foreign-language anxiety for decades.
It refers to anxiety specifically connected to using or learning another language: speaking in front of people, understanding fast speech, answering unexpectedly, being corrected, taking a language test, or worrying about how you sound.
That distinction matters.
Someone can be socially confident in their first language and still become tense when speaking English.
Likewise, a learner may feel comfortable doing grammar exercises, watching videos, or reading articles but become much more anxious when required to produce language in real time.
A large 2019 meta-analysis combined 55 independent samples with more than 10,000 participants and found a moderate negative relationship between foreign-language anxiety and language performance.
The overall correlation was about −.34.
That does not mean anxiety explains every bad conversation.
And correlation does not prove that anxiety alone causes lower performance.
Struggling with a language can itself create anxiety.
But across many studies, the two are meaningfully connected.
Speaking creates a special problem: there is no pause button
When you read English, you can stop.
When you write, you can delete a sentence and start again.
In a live conversation, the other person is waiting.
Your brain may need to do several things almost simultaneously:
- understand what the other person said;
- decide what you want to say;
- retrieve vocabulary;
- build grammatical structure;
- pronounce the sentence;
- monitor whether you are making mistakes;
- watch the listener's reaction;
- decide what to say next.
Fluent speakers automate much of this.
Language learners automate less of it.
That means a normal conversation can already consume a significant amount of attention.
Now add another process:
Do I sound stupid?
Was that grammar wrong?
Is my accent too strong?
Why are they looking at me like that?
I am taking too long.
Those thoughts are not linguistically useful.
But they still use attention.
Anxiety can compete with the system you use to speak
This idea has been investigated directly.
In a classic 1994 study, Peter MacIntyre and Robert Gardner examined anxiety at three stages of second-language processing: input, processing, and output.
They found that anxiety related to each stage was associated with poorer performance on tasks targeting that same stage.
The interesting part is the output stage.
You may correctly know an expression and still have more difficulty producing it when anxiety rises.
A modern systematic review published in 2025 looked specifically at foreign-language anxiety and cognition. Across the eligible studies, anxiety was generally associated with higher cognitive load, poorer attention, and weaker working-memory functioning.
That gives us a useful way to understand the familiar blank-mind moment.
The word may still be in memory.
Accessing it is simply becoming harder at the exact moment you need it.
The camera experiment makes this unusually concrete
One older experiment is especially relevant because researchers deliberately changed the social pressure of the task.
MacIntyre and Gardner had university students complete second-language vocabulary tasks.
At different stages, a video camera was introduced.
When the camera appeared, anxiety increased.
Performance also became worse.
The participants had not suddenly lost months of French study because a camera entered the room.
What changed was the psychological condition under which they had to use what they knew.
This distinction is important for learners who judge their entire English level from one uncomfortable conversation.
A stressful performance is still real performance.
But it may not reveal the full amount of language knowledge available under easier conditions.
Recent speaking research finds the same pattern in real speech measures
A 2024 study of Spanish learners of English examined speaking anxiety during oral narrative tasks.
The researchers measured several dimensions of spoken performance, including fluency, lexical and grammatical accuracy, pronunciation, accentedness, and comprehensibility.
More difficult tasks increased anxiety and harmed several aspects of performance.
Higher self-reported anxiety was associated with more breakdowns in fluency and lower lexico-grammatical accuracy.
Most strikingly, after controlling for participants' English proficiency and working-memory ability, anxiety still explained 13%–15% of the variation in breakdown fluency.
That is close to the experience learners often describe:
I know more English than it sounds like I know when I am nervous.
The research does not prove that this happens identically to every learner in every conversation.
But it makes the phenomenon much less mysterious.
Why do words disappear first?
Vocabulary retrieval is especially noticeable because failure is so obvious.
You know what you want to communicate.
You may even remember the first letter of the English word.
But nothing arrives quickly enough.
Word retrieval is not the same as storing a word in memory.
A word can be familiar and still be slow to retrieve.
A learner may therefore have several levels of access:
- recognize the word when reading;
- understand it when someone else says it;
- recall it with enough time;
- recall it quickly in a conversation;
- recall it quickly while also planning the next sentence and monitoring the listener.
Those are progressively harder conditions.
Anxiety does not need to erase the memory to cause trouble.
It only needs to make an already-demanding retrieval process less efficient.
This is also why the missing word often returns after the conversation.
The knowledge did not necessarily change.
The retrieval conditions did.
Monitoring yourself can become part of the problem
Some self-monitoring is useful.
You need to notice when a sentence is unclear. You sometimes need to repair grammar or pronunciation. You should care whether the listener understands you.
But anxious monitoring can become excessive.
Instead of using English to communicate, part of your attention starts watching yourself use English.
You hear your own accent.
You inspect your verb tense while the sentence is still leaving your mouth.
You notice a small pronunciation mistake and begin thinking about the mistake while trying to finish the next clause.
The conversation now has two tasks:
- communicate;
- evaluate yourself communicating.
For a learner with limited automaticity, that second task can be expensive.
Accent anxiety can change whether you speak at all
Pronunciation anxiety deserves special attention because learners often confuse having an accent with speaking badly.
These are not the same thing.
A 2017 qualitative study examined how learners' perceptions of accent affected their willingness to communicate in English.
Participants described themes including accent stigma, self-confidence, perceived communicative competence, and concern about how their accent would be received.
This matters because anxiety can influence language learning before a word is even spoken.
If you expect your accent to be judged, you may:
- give shorter answers;
- avoid unfamiliar words;
- choose safer sentence structures;
- speak less often;
- avoid conversations completely.
That creates an uncomfortable feedback loop.
You need real speaking experience to make speaking more automatic.
But anxiety can make you avoid the exact situations that would provide that experience.
Anxiety can reduce willingness to communicate
This is not only an accent issue.
A 2024 meta-analysis combined 50 independent studies examining second-language anxiety and willingness to communicate.
The relationship was again moderately negative: about r = −.34.
Learners with higher anxiety were generally less willing to initiate or enter communication in their second language.
This is one reason language anxiety can have effects beyond a single bad speaking performance.
If anxiety makes you speak less, you also receive less practice retrieving words under conversational conditions.
Less practice can preserve the gap between:
English you can recognize
and
English you can access instantly while another person is waiting.
The solution is not “stop being nervous”
That advice is almost useless.
If someone could switch anxiety off because they were told to relax, they probably would.
A better approach is to change the difficulty and stakes of the speaking situation while preserving actual communication.
Think of speaking pressure as something you can grade, not something that exists only as “safe” or “terrifying.”
For example:
Level 1: speak when nobody is listening
Describe your room aloud.
Retell a video.
Answer an imaginary question.
This helps convert passive vocabulary into retrievable speech, but it does not train social pressure.
Level 2: record yourself
Now there is a small sense of observation.
Do not restart every mistake.
The goal is finishing the message.
Level 3: send a voice message
You still have preparation time, but another person will actually hear you.
Level 4: have a short predictable conversation
Choose a familiar topic.
Order something. Ask one question. Explain something you know well.
Level 5: enter less predictable conversation
Now you have to listen, retrieve, adapt, and respond in real time.
The point is not that everyone must follow these exact five stages.
The point is to avoid making your only speaking practice either completely private or maximally stressful.
Practise retrieval under mild pressure
If your English disappears mainly in conversation, doing more recognition-only study may not fully solve the problem.
You also need practice retrieving language while attention is divided.
Try answering simple questions with a short time limit.
Give yourself ten seconds to begin speaking about:
What did you do yesterday?
or:
Which app do you use most and why?
Do not search for the perfect sentence.
Start with the language that is available.
This trains something flashcards and reading cannot perfectly reproduce: selecting and producing words before you have unlimited planning time.
Make recovery part of fluency
Anxious speakers often treat a missing word as a disaster.
Fluent communication treats it as a normal event.
You can continue without the exact word.
Instead of freezing, learn phrases such as:
I can't remember the exact word, but…
It's the thing you use for…
What I mean is…
Let me say that another way.
These phrases do not hide weak English.
They are communication tools.
Native speakers also forget words, restart sentences, explain around missing terms, and correct themselves.
The useful skill is not producing a flawless first attempt.
It is keeping the conversation alive when retrieval fails.
Do not make accent reduction the price of participation
Improving pronunciation can absolutely make speech easier to understand.
But “I should not speak until my accent sounds native” is a very different goal.
Accent is one characteristic of speech.
Comprehensibility is another.
Your practical target should usually be that listeners can understand you without unnecessary effort.
If an accent concern makes you avoid speaking, it can become more damaging than the accent itself.
You do not have to wait until your English sounds like somebody else's English before using it.
A bad conversation is data, not a language certificate
Suppose you speak badly in one conversation.
You pause constantly.
You forget easy words.
You leave thinking:
My English is terrible.
That conclusion contains too much information for one interaction to support.
A more useful diagnosis asks:
- Did I lack the vocabulary completely?
- Did I know the word but retrieve it too slowly?
- Was the topic unusually difficult?
- Was I trying to monitor every mistake?
- Was I worried about being judged?
- Did the same problem happen when I spoke alone?
Those are different problems.
And they require different solutions.
Studying another hundred words will not automatically fix a pressure-related retrieval problem.
Likewise, blaming anxiety for everything can hide genuine gaps in vocabulary or grammar.
You need to distinguish knowledge from access.
Is foreign-language anxiety permanent?
Probably not.
A 2025 systematic review of longitudinal studies examined whether foreign-language anxiety changes over time rather than behaving like a fixed personality trait.
The review treats language anxiety as dynamic and potentially malleable, while also noting that the research methods and time scales used across studies vary considerably.
That is a more useful model than thinking:
I am simply a nervous English speaker.
Your reaction can depend on the person, task, topic, setting, perceived evaluation, preparation, and previous experiences.
A learner who freezes during a formal presentation may speak comfortably with a friend five minutes later.
Same person.
Same English.
Different conditions.
So is your English actually worse when someone is listening?
Sometimes your performance probably is.
That is not imaginary.
Research connects foreign-language anxiety with lower performance, greater cognitive load, weaker attentional and working-memory functioning, reduced speaking fluency, and lower willingness to communicate.
But performance under pressure is not identical to stored language knowledge.
That distinction explains one of the most frustrating experiences in language learning:
You can know the English.
You can even prove that you know it in another context.
And still fail to access it quickly enough when somebody is waiting for you to speak.
The goal, then, is not only to learn more English.
It is to make the English you already know available under increasingly realistic conditions.
Practise retrieval.
Practise recovery.
Increase conversational pressure gradually.
And when one simple word disappears in the middle of a conversation, do not assume the word has vanished from your vocabulary.
Your brain may simply be busy doing too many things at once.
Sources and further reading
- Zhang — Foreign Language Anxiety and Foreign Language Performance: A Meta-Analysis, The Modern Language Journal
- MacIntyre & Gardner — The Subtle Effects of Language Anxiety on Cognitive Processing in the Second Language, Language Learning
- MacIntyre & Gardner — The Effects of Induced Anxiety on Three Stages of Cognitive Processing in Computerized Vocabulary Learning, Studies in Second Language Acquisition
- Foreign language anxiety and cognition: a systematic review and meta-analysis, The Language Learning Journal
- Mora, Mora-Plaza & Bermejo Miranda — Speaking anxiety and task complexity effects on second language speech, International Journal of Applied Linguistics
- Lou, Li & Zhao — A Meta-analysis of the Relationship Between L2 Willingness to Communicate and L2 Anxiety
- Zarrinabadi & Khodarahmi — L2 Willingness to Communicate and Perceived Accent Strength: A Qualitative Inquiry
- Is foreign language anxiety malleable? A systematic review of longitudinal observation studies, System