Why Known Words Get Stuck on the Tip of Your Tongue

A tip-of-the-tongue state is a failure to retrieve the sound of a word whose meaning you still have. Second-language speakers get them more often because each word is used less, so the link from meaning to pronunciation is weaker. The fix is to learn and rehearse the sound, not to reread the definition.

What is actually happening when a word gets stuck

You know the word. You could pick it out of a list. You might even know it starts with an s and has two syllables. It just will not come out. That specific, maddening state has a name in memory research, and it has been studied for sixty years, so the picture of what is going on is unusually clear.

The founding study is Brown and McNeill (1966). They read college students definitions of rare English words, “to give up the throne” for abdicate, “a navigational instrument for measuring the angle of the sun or stars” for sextant, and waited for the students to get stuck. Several hundred did. And while stuck, the students could report things they had no business knowing if the word were really gone: the first letter, the number of syllables, where the stress fell, and words that sounded similar. Someone hunting for sextant offered secant and sextet.

That partial knowledge is the whole point. In a tip-of-the-tongue state the word is not missing. The entry is there, the meaning is there, and a good part of the form is there. What fails is the last step, assembling the full sound and getting it out.

The meaning survives, the sound fails

Speaking a word takes two retrievals, not one. First you find the word as an idea, the thing that means “give up the throne”. Then you find its pronunciation. Those are separate lookups, and they can fail independently.

Burke, MacKay, Worthley and Wade (1991) built the standard account of why the second one is the fragile one. In their model, a word’s meaning is connected to its sound by links that weaken when the word has not been used often, or not used recently. Meaning has many redundant connections, so it survives a weak link here and there. Sound has far fewer, close to one route per word, so when that route is under-primed nothing arrives at the other end. You get the concept in full and the pronunciation in fragments.

This is why a tip-of-the-tongue state feels the way it does. It is not vagueness. You know exactly which word you want. What you have lost, temporarily, is the path from that word to your mouth.

For a language learner this has an uncomfortable implication. Rereading a definition strengthens the side that never broke.

Why it happens more in a second language

Everyone gets tip-of-the-tongue states. People who speak two languages get more of them, and the reason turns out to be simpler than “two languages interfere”.

Gollan and Silverberg (2001) first showed that Hebrew-English bilinguals had more tip-of-the-tongue states than English monolinguals of the same age. Gollan and Acenas (2004) then ran the study that explained it. They showed Spanish-English and Tagalog-English bilinguals pictures and asked them to name the object in English. Bilinguals got stuck more often than monolinguals, with one exception. When the English word was a cognate, like vampire and Spanish vampiro, the gap closed. For non-cognates, like funnel and embudo, it stayed open.

Their explanation is called the weaker links hypothesis, and it is the most useful idea in this article. If you speak two languages, your speaking time is split between them. Every word in each language gets used less than it would if you had only one. Less use means weaker links from meaning to sound, and weaker links mean more words that stall on the way out. Cognates escape, on the usual reading, because saying the Spanish word partly exercises the English one.

Two later findings nail the explanation down. Gollan, Montoya and Bonanni (2005) found that for proper names, bilinguals and monolinguals got stuck equally often. Names are rare for everybody, so there was no use gap to widen. And Pyers, Gollan and Emmorey (2009) found the same excess of tip-of-the-tongue states in people bilingual in English and American Sign Language, whose two languages share no sounds at all. So two sound systems colliding cannot be the cause. What is left is use. Each word simply gets fewer reps.

Read that as a learner and the diagnosis changes. A word getting stuck says nothing about how good your English is. It says how many times you have said that particular word out loud.

Which words get stuck

The weaker links account makes a prediction about which words will stall, and it matches what learners report.

The word	Why it stalls	What protects it
Low-frequency words	Few uses, so the meaning-to-sound link never got strong	Deliberate spoken retrieval
Words learned from reading alone	There was never a stable sound to retrieve in the first place	Learn the pronunciation with the word
Words you have not said in months	The link has decayed since the last use	Spaced review that reaches speech
Long words with unusual stress	More sound to assemble, more places to fail	Rehearse the stress pattern aloud
Non-cognates	No help from your first language	Extra spoken reps
Proper names	Rare for everyone, one name per referent	Not much. They stall for native speakers too

The second row is the one that matters most for learners, and it is the least discussed. If you learned a word from a page and never heard it, then the “sound” your brain filed is a guess assembled from the spelling. English spelling being what it is, the guess is often wrong, and even when it is right it is thin. The word will look fine on the page and refuse to come out in conversation, and you will blame your memory when the real problem is that nothing was ever stored in the slot the retrieval needs.

Struggling makes it worse

The instinct when a word is stuck is to hold on and grind. Warriner and Humphreys (2008), in the Quarterly Journal of Experimental Psychology, showed that this is the wrong move.

They induced tip-of-the-tongue states with definitions, and each time a participant reported being stuck, they were randomly given either 10 or 30 seconds to keep trying before being told the answer. Two days later everyone was tested on the same words. A word that had produced a tip-of-the-tongue state in the first session was more than four times as likely to produce one again. And the longer the participant had been left to struggle the first time, the more likely the word was to stick again.

The authors’ interpretation is what gives the paper its title, “Learning to fail”. Time spent in the failed state is practice. You are rehearsing the wrong retrieval path, and the brain learns it like anything else.

A 2015 follow-up by D’Angelo and Humphreys, in Cognition, added the useful half. The error learning was corrected when the stuck word got resolved, whether it came back on its own or with a cue. Getting to the word yourself, rather than being stranded in the gap and then handed the answer, is what breaks the pattern.

Put the two together and the rule is simple. Give a stuck word a few seconds, and if it is not coming, get to it by any means available rather than sitting in the gap.

What to do in the moment

Since the partial information is real, it is usable. A stuck word can usually be fished out with cues that target the sound side, which is where the break is.

• Say the first letter or sound you have, out loud. Then try a few words that start with it.
• Count the syllables you feel it has and try to hum the stress pattern. Brown and McNeill’s participants knew both more often than chance would allow.
• Run through the words it sounds like. The near-misses are often one sound away from the target.
• If none of that works within about ten seconds, paraphrase and keep talking. “The instrument sailors used to measure the sun” gets the conversation through.
• Look the word up as soon as the conversation allows. Do not leave it for later. Every extra minute in the gap is a rep on the failure.
• When you find it, say it out loud, in a sentence, three times. The repair has to happen on the sound side because that is where the break was. Reading it silently repairs nothing.

That last point is the one people skip, and the shadowing research is fairly clear on why it costs them. A word that gets spoken during study gets a distinct sound record that silent reading does not create.

How to study so words stop getting stuck

Everything above points at the same conclusion. The words that stall are the ones whose sound was never learned or never maintained. So the study routine has to build the sound and then keep it.

Learn the pronunciation with the word, on day one. Not later, not once you know it. A word entered into memory as spelling plus meaning has an empty sound slot, and everything downstream of that is a retrieval failure waiting to happen. Read the IPA, listen to the audio, and say it before you move on. This is the single change with the largest effect on tip-of-the-tongue states, because it fixes the row in the table that learners fall into most.

Retrieve in the speaking direction. Most review shows you the word and asks for the meaning. That practices comprehension. To practice speech you need the other direction. See the meaning, or a picture, or a sentence with a gap, and produce the word. This is the retrieval that fails in conversation, so it is the one that needs the reps.

Say it aloud when you review. MacLeod and colleagues (2010) found that words read aloud during study were remembered better than words read silently, by a margin of roughly 10 to 25 percent across their experiments. That was list memory in a lab, not conversation, so treat the number as a direction rather than a promise. The direction is the right one.

Keep the links from decaying. The weaker links account says links weaken with time since last use. That is a scheduling problem, and it is the problem spaced repetition exists to solve. A scheduler like FSRS estimates when a word’s recall is about to slip and puts it in front of you just before, so each review is a use that arrives in time. Words you keep retrieving cleanly drift out to longer intervals. Words that stall come back sooner.

Use it in speech within a day. A word that has been retrieved out loud in a real sentence, to a real person, has had the kind of rep the research says it needs. A single real conversation does more for that word than another silent pass through the deck.

Where this leaves your routine

A tip-of-the-tongue state is a diagnostic. The word’s sound was learned too thinly or used too rarely, and the stall tells you which word. The moment it happens, get out fast and say the word. Over the longer run, learn every word with its pronunciation, retrieve it in the speaking direction, and let a schedule keep the link alive.

Vocaby is built around that sequence. Every one of its 29,000+ English words comes with IPA and audio, a plain-English definition and real example sentences, so the sound goes in with the meaning instead of being reconstructed from the spelling later. FSRS handles the timing, so the words that tend to stall are the ones that come back soonest. Browse the collection at our word lists.

The word is in there. Give it a sound, and it will come out.

Learn English vocabulary with Vocaby on the App Store