The phonological loop is the part of your working memory that temporarily holds verbal and sound-based information. It’s what lets you repeat a phone number in your head long enough to dial it, silently rehearse a grocery list while walking through the store, or hold the beginning of a sentence in mind while you read to the end. First proposed by psychologists Alan Baddeley and Graham Hitch in 1974 as one component of their broader working memory model, the phonological loop has become one of the most studied structures in cognitive psychology.
How the Phonological Loop Works
The phonological loop has two parts that work together. The first is a passive storage buffer, sometimes called the “inner ear.” This is where sound-based memory traces land and sit briefly. The catch is that these traces fade fast, within about one to two seconds, unless something refreshes them.
That refreshing is handled by the second component: an active rehearsal process often described as the “inner voice.” This is the subvocal speech you use when you silently repeat something to yourself. As long as you keep rehearsing, the memory trace stays alive. Stop rehearsing, or get interrupted, and the trace decays.
Even visual information can enter the loop. When children aged seven and older are shown pictures, they tend to silently name what they see, converting the visual input into a sound-based code they can rehearse. Adults do this constantly. You don’t just see the word “cat” on a page. You hear it internally, and that sound-based version is what the phonological loop holds onto.
What Limits Its Capacity
The phonological loop doesn’t have a fixed number of “slots.” Its capacity depends on how quickly you can rehearse material before it decays. This creates a word length effect: you can remember more short words than long ones, because short words cycle through the rehearsal loop faster. In classic experiments by Baddeley and colleagues, recall performance dropped steadily as the number of syllables in each word increased. Even among two-syllable words, shorter-duration words (around 70.7% recall accuracy) were remembered better than longer-duration ones (65.5% accuracy).
Interestingly, the effect isn’t purely about how long a word takes to say out loud. Research comparing words matched for spoken duration but differing in syllable count found that words with fewer syllables were still easier to recall. The complexity of planning the mouth movements for a word, not just its duration, appears to play a role in how much rehearsal effort it demands.
What Disrupts It
Two well-documented interference effects reveal a lot about how the loop operates. The first is the phonological similarity effect: words that sound alike are harder to remember in sequence. If you try to hold “man, cap, cat, map, mad” in memory, you’ll do worse than with a list of dissimilar-sounding words, because the similar sound patterns blur together in the phonological store.
The second is the irrelevant sound effect. Background speech, even in a language you don’t understand, disrupts the phonological loop and increases errors during recall tasks. This happens because spoken sounds gain automatic access to the phonological store whether you want them there or not. It’s why trying to memorize something in a noisy café feels so much harder than doing it in a quiet room.
There’s also articulatory suppression: if you’re asked to repeat a meaningless syllable like “the, the, the” while trying to remember a word list, your rehearsal mechanism is occupied and can’t refresh the memory traces. The phonological store still functions, but without rehearsal, the traces become much more vulnerable to interference and decay.
Where It Lives in the Brain
Brain imaging studies have mapped the two components of the phonological loop to distinct regions in the left hemisphere. The passive storage component is associated with the inferior parietal lobe, specifically a region called the supramarginal gyrus. This area appears particularly important for maintaining the order of items in a sequence, not just the items themselves.
The rehearsal component activates areas in the left inferior frontal region (overlapping with Broca’s area, known for speech production) and the left lateral premotor cortex. This makes sense: subvocal rehearsal is essentially silent speech, so it recruits the same neural machinery used for planning and producing spoken language. The two regions are connected by a white matter pathway, and when that connection is disrupted during neurosurgery, patients make significantly more errors on digit span tasks.
The Phonological Loop in Everyday Life
You rely on the phonological loop more than you probably realize. Hearing a phone number on a radio ad and holding it in mind while you grab your phone is a textbook example. But the loop also supports reading comprehension (holding the first half of a long sentence while you process the second half), mental arithmetic (keeping intermediate numbers in mind), following multi-step directions, and learning new vocabulary.
In children learning to read, the phonological loop plays a critical role in sounding out unfamiliar words. A child decoding the word “attempt” needs to hold the sounds /a/ and /tempt/ in the loop long enough to blend them together. When the loop struggles, the child might recall only the first sounds and guess the rest, reading “attempt” as “attic.” Similarly, a beginning reader might successfully sound out /k/ /a/ /t/ but then say “kitty” instead of “cat,” because the phonological traces partially decayed before blending was complete.
Connections to Dyslexia and Language Disorders
Weaknesses in the phonological loop are closely linked to both dyslexia and specific language impairment (SLI). Children with dyslexia consistently perform poorly on tasks that tax the loop, such as digit span (repeating back a string of numbers) and nonword repetition (repeating made-up words like “blonterstaping”). These tasks require holding unfamiliar sound sequences in the phonological store and rehearsing them accurately, exactly the skills the loop provides.
In SLI, poor nonword repetition is one of the most reliable behavioral markers, reflecting deficits in both phonological storage and verbal short-term memory. One prominent theory holds that this phonological loop weakness cascades into broader language problems: if you can’t hold new sound patterns in memory long enough, learning new words and processing complex sentences becomes much harder. Phonological skills, including the loop’s storage and rehearsal functions, are also strongly predictive of how well children acquire reading ability.
How the Model Has Evolved
The phonological loop has remained a stable part of cognitive theory for over fifty years, which is unusual in psychology. The broader working memory model has seen relatively few major revisions. The most notable came in 2000, when Baddeley added a new component called the episodic buffer, a system for integrating information across different memory stores. But the phonological loop itself has held up well.
That said, it’s not without challenges. The original model described the loop’s behavior in broad terms without specifying exactly how serial order information is encoded and maintained. More recent work has also shown that some effects once thought to be unique to the phonological loop, like sensitivity to word length and phonological similarity, also appear in other memory tasks that weren’t originally thought to involve the loop. These findings suggest the boundaries between memory systems may be less sharp than the original model implied, but the core idea of a sound-based rehearsal system with rapid decay remains well supported.