The Phonological Loop — How Your Brain Rehearses Sound

Written by TOMOY · Published: April 8, 2026 · Updated: April 8, 2026

1. What Is the Phonological Loop?

The phonological loop is one of the subsystems in Baddeley and Hitch's model of working memory. It is responsible for holding verbal and auditory information in a temporary, accessible form — the kind of mental process engaged when you silently repeat a phone number to yourself while searching for a pen.

Unlike long-term memory, the phonological loop operates over very short timescales. Without active maintenance, information held in it fades within a few seconds. What keeps it alive is a process of internal repetition — the brain's equivalent of speaking the items again before they decay.

2. Two Components: Store and Rehearsal

Baddeley's model describes the phonological loop as having two distinct parts:

Phonological Loop Structure

Phonological Store A passive buffer that holds speech-based representations for approximately 1–2 seconds before they decay. Any spoken or heard information enters here automatically.

Articulatory Rehearsal Process An active, speech-like process that refreshes items in the phonological store by "saying" them internally. Also converts written words into a phonological code so they can enter the store.

These two components work together: the store provides temporary holding space, and the rehearsal process keeps items from decaying by continuously refreshing them. The loop can be thought of as a mental tape recorder running on a short loop.

3. The Word Length Effect

One of the clearest pieces of evidence for the phonological loop comes from research on the word length effect: people tend to recall fewer long words than short words from a list, even when the lists contain the same number of items.

The explanation offered in the literature is that the articulatory rehearsal process takes longer for longer words. If rehearsal takes more time per item, fewer items can be refreshed before they decay in the phonological store — and the effective capacity of the loop shrinks accordingly. Research has found that people tend to recall approximately as many words as they can read aloud in about two seconds, suggesting the capacity of the loop is tied to rehearsal speed rather than a fixed item count.

4. The Phonological Similarity Effect

A related phenomenon is the phonological similarity effect: sequences of words that sound alike (such as "cat, bat, mat, rat") are typically harder to recall correctly than sequences of dissimilar-sounding words. The phonological store holds items as sound-based representations, so when multiple items share similar phonological features, their traces in the store become harder to distinguish from one another.

This effect is notably absent for visual presentation of words — as long as the reader does not sub-vocalize while reading, visually presented sequences of rhyming words are only marginally more difficult than non-rhyming ones. This dissociation supports the view that the phonological store is specifically tied to sound-based representations, not visual ones.

5. Digit Span Tasks and the Phonological Loop

Digit span tasks — in which a sequence of digits must be recalled in order — place heavy demand on the phonological loop. Digits are typically short, phonologically distinct items, which makes them well-suited to probing the store's capacity under controlled conditions. This is part of why digit span has been used as a measure of immediate verbal memory in cognitive psychology research for many decades.

When digit sequences are reversed or reordered (as in backward span or ascending span tasks), the demand shifts toward the central executive, which must manipulate the contents of the phonological store rather than simply hold them. This is why those variants are considered more cognitively demanding than forward recall: they require engagement of an additional system beyond the loop itself.

6. Limitations of the Phonological Loop

The phonological loop is a domain-specific resource. It is engaged by verbal and auditory material but does not handle visual or spatial information directly — that is the role of the visuospatial sketchpad. When both verbal and visual demands are present simultaneously, they draw on separate subsystems and can, to a degree, operate in parallel without interference.

The loop also has no direct access to semantic memory — it holds sounds, not meanings. A string of nonsense syllables occupies phonological loop capacity just as a meaningful sentence does. This is why strategies like chunking (grouping items by meaning) primarily involve the central executive and long-term memory, not the loop itself.