← All articles
psychologymemorywordsspeechSeptember 17, 20263 min read

Why Can You Almost Remember a Name? The Meaning Arrived Without the Sound

By the BrainSnail editorial team. How these articles are written and checked, and how to tell us when one is wrong.

Knowing exactly what you want to say while being unable to produce the word is a distinct and well-studied state, in which partial information about the word is available and the word itself is not. What people recall in that state is revealing.

What is actually available

The experience feels like having nothing and experiments show that a substantial amount is present. People in this state can frequently report the first letter of the word at above chance rates, can estimate the number of syllables correctly, can identify where the stress falls, and can reject candidate words as wrong with confidence while being unable to produce the right one. They also recall words that sound similar and words that mean similar things, which they know are not what they want. That pattern shows that the meaning has been retrieved successfully and the sound has not, which is the central finding and which supports a model in which those are stored and accessed separately.

What triggers it

The state occurs far more for some words than others:

  • Proper names, especially people's names, by a large margin
  • Words encountered rarely
  • Words not used recently even if known well
  • Words with few similar-sounding neighbours in the language
  • More often in older people, which is a robust and well-documented finding
  • More often in bilingual speakers, for each of their languages

Why names are the worst

Personal names are retrieved less reliably than any other category, and the explanation concerns what a name connects to. An ordinary word is linked to a meaning that connects to many other concepts, so there are many routes to reach it, and activation spreading from any of them can retrieve it. A name is arbitrary, since nothing about a person determines what they are called, so it connects to the person and to nothing else, which leaves exactly one route. Names are also frequently unique rather than shared, so no other knowledge reinforces them. That explains why somebody can recall a person's job, their town, their family and where they met them while the name itself remains unavailable.

How it was studied

Investigating a state that arrives unpredictably requires producing it on demand, and the method devised for that is elegant. Roger Brown and David McNeill published a study in 1966 in which they read participants the definitions of uncommon words and asked what the word was, which produced the state reliably in a proportion of cases, and then questioned people in that state about what they could report. That design established the whole field and is still used. Later work added brain imaging, showing activity in regions associated with detecting a conflict and with searching rather than with retrieving, which fits the subjective experience of knowing that something is missing and looking for it.

What to do about it

The evidence about recovery is clear and contradicts the usual advice. Searching actively frequently makes it worse, because generating candidate words that are close but wrong strengthens those candidates and blocks the target further, an effect demonstrated experimentally by giving people related words and showing that it reduces retrieval. Stopping and returning later works far better, and the word frequently arrives unbidden minutes or hours afterwards. Being given the first sound resolves it immediately in most cases, which is why asking somebody else for a hint works. Going through the alphabet is the folk method and has modest support. And the state resolves eventually in the great majority of cases without anybody doing anything.

The takeaway

People in this state can report the first letter, the syllable count and the stress pattern above chance while being unable to produce the word, which shows meaning is retrieved and sound is not. Personal names suffer most because a name connects to the person and to nothing else, leaving one route. Searching actively strengthens the wrong candidates and blocks the target, and stopping works better.

Practise this

Questions from Memory

Reading about something is not the same as being able to recall it. These are real questions from the Memory unit in our Psychology track, answers and explanations included. The unit has 120 in total across 23 steps.

  • Multiple choiceLevel 2

    1. Which statement about memory is most accurate?

    • The brain rebuilds a memory each time we recall itcorrect
    • Every memory is a perfect copy of the event
    • Memories can never change once they are made
    • We remember every detail of our whole life

    Memory is reconstructive, so the brain pieces a memory back together and can add or change details.

  • Multiple choiceLevel 1

    2. What are the three basic jobs your memory does with information?

    • Encoding, storage, and retrieval
    • Reading, writing, and math
    • Seeing, hearing, and touching
    • Sleeping, eating, and playing

    Memory works by encoding information in, storing it over time, and retrieving it back out.

  • True or falseLevel 2

    3. Reviewing something a little each day helps you remember it better than cramming it all at once.

    Answer: True

    Spacing your practice over several days builds stronger, longer lasting memories than cramming.