Proactive Inhibition

Abstract

Proactive inhibition is a form of psychological inhibition in which earlier learning impairs the retention of material acquired afterward. Benton Underwood showed that most forgetting of verbal material reflects interference from prior lists rather than passive decay, and the classical two-factor theory attributes the loss to unlearning of old associations during new learning and to competition among responses at recall. The Keppel and Underwood demonstration that short-term forgetting itself grows with prior trials, and the Wickens release-from-inhibition paradigm, made the buildup measurable. Modern work recasts proactive inhibition as a failure of executive control in working memory. Three interactive demonstrations trace the buildup and release of inhibition, its dependence on prior learning, and its signature in working-memory recognition.

Keywords: proactive inhibition, interference, forgetting, verbal learning, working memory

Proactive inhibition, also called proactive interference, is the phenomenon in which previously learned material intrudes upon and degrades the retention of material learned later. The term is the older name, drawn from the verbal-learning tradition, for what much of the contemporary literature calls proactive interference; Medical Subject Headings retains proactive inhibition as the preferred descriptor, and the two labels denote the same effect. Its discovery reversed the intuitive picture of forgetting. If memories simply faded with time, then what a person learned before studying a new list should be irrelevant to how much of that new list survives. Underwood's analysis of decades of retention data showed the opposite: the more lists a person had learned beforehand, the less of the current list they retained a day later, so the prior learning, not the passage of time, was doing most of the damage (Underwood, 1957).

What Proactive Inhibition Is

Proactive inhibition is defined by the direction of the interference: old learning reaching forward to disrupt new. A person who has memorised several lists of paired associates and then learns one more will recall the final list more poorly than a person who learned only that list, even when study time, list length, and the retention interval are matched. The only difference is the interfering history the two people bring to the task, and that history is what the term names. It is distinguished sharply from retroactive interference, in which the later learning reaches back to impair the earlier; the two are complementary halves of interference theory, and both must be controlled to estimate how much a memory has genuinely decayed (Underwood, 1957).

The historical importance of the effect lies in what it displaced. Early in the twentieth century, forgetting was widely modelled as the passive fading of a trace, so that retention was a function of elapsed time alone. Underwood re-examined the many studies reporting that participants forgot roughly three-quarters of a learned list within a day and noticed that the participants in those studies were typically seasoned subjects who had learned list after list. When he plotted forgetting against the number of prior lists, the supposed universal forgetting rate dissolved into a family of curves: a person learning their first list forgot very little, and only those with long interfering histories forgot most of what they had studied (Underwood, 1957). The name inhibition reflects the era's assumption that the prior material actively suppresses the newer, an assumption the two-factor theory then tried to make mechanistic.

The Two-Factor Theory: Unlearning and Competition

The classical account of interference, consolidated by Postman and Underwood, proposes two mechanisms acting together (Postman & Underwood, 1973). The first is unlearning: when a person learns a new response to an old stimulus, the earlier association is weakened, much as an extinguished conditioned response is suppressed. The second is response competition: at the moment of recall, both the old and the new responses are elicited by the cue and compete for output, so the correct response may lose the race even if it survives in storage. Unlearning reduces the availability of a response; competition reduces its accessibility given that it is available.

The two factors make different predictions across the retention interval, and this is where proactive inhibition earns its name. Immediately after learning, the newest associations dominate and competition favours the correct current response, so proactive inhibition is small. Over minutes and hours, the temporarily suppressed older associations spontaneously recover strength, and as they do they compete more effectively against the current response, so proactive inhibition grows with delay. The prediction that forgetting should increase over the retention interval specifically because old material recovers, rather than because new material decays, is the theory's signature, and it was confirmed most cleanly in the short-term paradigm described next (Keppel & Underwood, 1962).

Release From Proactive Inhibition

The most influential experimental handle on proactive inhibition came from the short-term retention task of Peterson and Peterson, in which a person holds a single item, such as a consonant trigram, across a brief filled delay and then recalls it (Peterson & Peterson, 1959). The rapid forgetting observed in that task was first read as decay of the short-term trace. Keppel and Underwood showed that it was largely proactive inhibition: on the very first trial, when no prior items could interfere, retention across the delay was nearly perfect, and the steep forgetting curve appeared only from the second and third trials onward, as earlier trigrams accumulated to interfere with the current one (Keppel & Underwood, 1962).

Wickens turned this buildup into a precise instrument. If the interference comes from earlier items that share some encoded property with the current one, then changing that property on a later trial should abolish the interference and restore recall. Wickens, Born, and Allen presented several trials of trigrams drawn from one class, watched recall decline as inhibition built up, and then switched to a class differing on some dimension, such as digits after consonants or one taxonomic category after another (Wickens et al., 1963). Recall rebounded sharply on the shift trial, a rebound called release from proactive inhibition. Because the size of the release measures how different the encoding system treats the new class from the old, Wickens used it as an assay of the dimensions along which words are semantically encoded, finding large releases for shifts of meaning and smaller ones for shifts of surface form (Wickens, 1970). The demonstration below reproduces the buildup across same-category trials and the release produced by a category shift.

Determinants: Similarity, Prior Learning, and Time

Three variables govern the magnitude of proactive inhibition, and each maps onto one of the mechanisms above. The amount of prior learning is the variable Underwood isolated: recall of the current material falls as the number of previously learned lists rises, because more prior associations are available to compete (Underwood, 1957). The similarity between the old and the new material is the variable Wickens exploited: interference is greatest when the two share an encoding dimension and is released when they diverge, so a category shift restores recall (Wickens, 1970). The retention interval is the variable the two-factor theory foregrounds: because suppressed older associations recover over time, proactive inhibition is negligible immediately after study and substantial after a delay (Keppel & Underwood, 1962). The demonstration below traces the first of these, the growth of forgetting as the interfering history lengthens, in the form Underwood originally reported.

Note. Table 1. The three classical determinants of proactive inhibition, each aligned with a mechanism of the two-factor theory, together with the individual-difference variable added by later working-memory research.

Proactive Inhibition in Working Memory and Executive Control

The interference tradition faded from fashion in the 1970s but returned, recast in the vocabulary of working memory and executive control. Kane and Engle showed that susceptibility to proactive interference is not uniform: people with lower working-memory capacity accumulate more interference across trials and suffer more when attention is divided, implying that resisting proactive inhibition is an act of controlled attention rather than a passive property of the trace (Kane & Engle, 2000). Anderson pressed the reframing further, arguing that the phenomena interference theory catalogued, including the unlearning that classical theorists posited, are produced by an active inhibitory process under executive control that suppresses competing memories during retrieval (Anderson, 2003).

Neuroimaging localised part of this control. Using a recognition task in which a probe item matches something held on a recent but now-irrelevant trial, Jonides and Nee found that resolving the resulting proactive interference recruits the left inferior frontal gyrus of the prefrontal cortex, the same region implicated in selecting among competing representations (Jonides & Nee, 2006). In such tasks a recently seen item that is no longer correct produces a slow, error-prone rejection, and the more recent the interfering item, the stronger the false familiarity that must be overridden. The demonstration below models this signature: the probability of a false recognition as a function of how recently the interfering item appeared.

Worked Example

Underwood's (1957) central claim can be made quantitative with a simple decay-of-recall-with-history model in the spirit of his own figures. Let the proportion of a freshly learned list still recalled after twenty-four hours be a declining function of the number of lists learned beforehand, R(n) = 18 + 62 e^(-0.12 n), where n is the number of prior lists and R is a percentage. The constant 18 is the asymptotic retention floor for a heavily practised subject, and 80 (the sum 18 + 62) is the retention of a person learning their very first list (Underwood, 1957).

- No interfering history. With n = 0, R = 18 + 62 e^(0) = 18 + 62 = 80.0%. A person learning their first list forgets only a fifth of it overnight. - A moderate history. With n = 5, R = 18 + 62 e^(-0.6) = 18 + 62(0.5488) = 18 + 34.0 = 52.0%. Five prior lists nearly double the overnight loss. - A long history. With n = 20, R = 18 + 62 e^(-2.4) = 18 + 62(0.0907) = 18 + 5.6 = 23.6%. The seasoned subject retains under a quarter of the new list.

The model makes Underwood's point concrete: the difference between recalling 80% and recalling 24% of the very same list is due entirely to the interfering history, not to any property of the list or the delay. The single supposed forgetting rate of the older literature was an average over people with different histories, and once history is entered as a variable the rate is revealed as a family of curves. The demonstration above plots this function and lets the interfering history be varied directly.

Discussion

Proactive inhibition reshaped the theory of forgetting by showing that a memory's fate depends on what surrounds it, not merely on how long it must last. Underwood's demonstration that prior learning, rather than time, drives most laboratory forgetting made interference the dominant account of retention loss for a generation, and the two-factor theory of unlearning and competition supplied its mechanism (Underwood, 1957; Postman & Underwood, 1973). The release-from-inhibition paradigm converted the phenomenon from a nuisance into a measuring instrument, one sensitive enough to chart the dimensions along which words are encoded (Wickens, 1970). The later migration of the construct into working memory did not overturn these findings so much as redescribe them: the buildup of interference across trials and its dependence on attentional resources are now read as the cost of controlling which of several competing memories reaches output (Kane & Engle, 2000; Anderson, 2003). What remains contested is whether a single inhibitory control process underlies the classical unlearning, the release effect, and the working-memory findings alike, or whether the shared label conceals distinct mechanisms that happen to converge at the point of retrieval.

Current Directions

Recent work has pushed on the boundary between interference and the capacity limits of memory itself. Endress and Szabó argued that several apparent capacity limits, including the rapid loss of items in short-term memory, are better understood as consequences of interference among similar representations than as fixed storage bounds, tightening the link between proactive inhibition and the architecture of working memory (Endress & Szabó, 2017). Kliegl and Bäuml synthesised the cognitive and neural evidence on the buildup of, and release from, proactive interference, drawing the behavioural paradigm together with its candidate prefrontal and hippocampal substrates (Kliegl & Bäuml, 2021). A parallel strand asks how long-term memory shapes the interference seen in working memory across the lifespan: Rhodes, Buchsbaum, and Hasher found that older adults show both greater proactive interference and, under some conditions, greater proactive facilitation from prior knowledge, suggesting that the same reliance on long-term memory can help or hurt depending on the match between old and new (Rhodes et al., 2022). Most recently, Mercer showed that the familiarity of the interfering material modulates proactive interference in verbal memory, reconnecting the modern paradigm to the similarity principle Wickens established six decades earlier (Mercer, 2025).

Commonly Confused With

Retroactive Interference

The prefix names what does the interfering, not what gets harmed. Pro- means the older learning reaches forward to disrupt the newer; retro- means the newer learning reaches back to disrupt the older. Ask which one is the culprit: in proactive inhibition the earlier material is, in retroactive inhibition the later material is. The confusion is durable because the direction is not in the everyday sense of the words, and because the two effects operate together and each is itself an instance of interference.

Proactive Interference

Not a different phenomenon, but the same one under a newer name. The verbal-learning literature coined proactive inhibition, on the assumption that old material actively suppresses new; the working-memory literature prefers proactive interference, which is theory-neutral about the mechanism. Medical Subject Headings indexes the effect under proactive inhibition, so the two terms are used interchangeably here, with interference preferred whenever no suppressive mechanism is being claimed.

Common Misconceptions

Forgetting is mainly the fading of memories over time.

Most laboratory forgetting of verbal material reflects interference from other learning, not passive decay. Underwood showed that a person learning their first list forgets little overnight, while the heavy forgetting reported in earlier studies occurred only in subjects with long histories of prior lists (Underwood, 1957).

Rapid forgetting in short-term memory proves that the trace simply decays.

The steep short-term forgetting curve is largely proactive inhibition. On the first trial of a Peterson-style task, when nothing has come before, retention across the delay is nearly perfect; the curve appears only as earlier items accumulate to interfere (Keppel & Underwood, 1962).

Proactive inhibition is a fixed property of the memory trace.

Susceptibility to interference varies with a person's attentional control. Those with lower working-memory capacity accumulate more proactive interference and suffer more under divided attention, so resisting it is an act of executive control, not a passive feature of storage (Kane & Engle, 2000).

Glossary

Accessibility.

Whether an available memory can actually be retrieved under the current cue conditions; response competition reduces accessibility.

Availability.

Whether a memory trace exists in storage, independent of whether it can be retrieved on a given test; unlearning reduces availability.

Buildup of inhibition.

The progressive decline in recall across successive trials that share an encoding dimension, as interference from earlier items accumulates.

Decay theory.

The account that memories fade as a function of elapsed time alone; largely displaced for verbal material by interference theory.

Executive control.

The set of attentional processes that select among competing representations; recruited to resolve proactive interference in working memory.

Interference theory.

The account of forgetting on which memories impair one another rather than fading in isolation, comprising proactive and retroactive interference.

Paired-associate learning.

A verbal-learning task in which a stimulus term must be linked to a response term, the classic arena for studying interference between successive lists.

Proactive inhibition.

The impairment of retention of later-learned material by earlier learning; the subject of this article, also called proactive interference.

Release from proactive inhibition.

The rebound of recall on a trial whose material shifts to a new encoding class, used to index the dimensions along which items are encoded.

Response competition.

The interference mechanism by which old and new responses elicited by the same cue compete at output, so the correct one may lose the race.

Retention interval.

The time between study and test; proactive inhibition grows across this interval as suppressed older associations recover strength.

Retroactive interference.

The mirror image of proactive inhibition, in which later learning impairs the retention of earlier learning.

Two-factor theory.

The classical account attributing interference to unlearning of old associations during new learning and to response competition at recall.

Unlearning.

The weakening of an earlier stimulus-response association when a new response is learned to the same stimulus, one of the two factors in interference theory.

Working memory.

The limited-capacity system that maintains and manipulates information over seconds; its capacity predicts susceptibility to proactive interference.

Key Researchers

Michael C. Anderson (b. 1962). Programme leader at the MRC Cognition and Brain Sciences Unit, University of Cambridge; he recast interference as an active, executive inhibitory process that suppresses competing memories during retrieval. ORCID - Google Scholar - Faculty Page

Karl-Heinz T. Bäuml. Professor of Experimental Psychology at the University of Regensburg; with Kliegl he synthesised the cognitive and neural mechanisms of the buildup of, and release from, proactive interference. ORCID - Google Scholar - Faculty Page

John Jonides (b. 1944). Professor of Psychology at the University of Michigan; with Nee he localised the resolution of proactive interference in working memory to the left inferior frontal gyrus. ORCID - Google Scholar - Wikipedia - Faculty Page

Michael J. Kane (b. 1968). Professor of Psychology at the University of North Carolina at Greensboro; with Engle he tied individual differences in working-memory capacity to susceptibility to proactive interference. ORCID - Google Scholar - Faculty Page

Benton J. Underwood (1915-1994). Professor of Psychology at Northwestern University; he demonstrated that proactive interference from prior learning, not decay, accounts for most laboratory forgetting of verbal material, founding modern interference theory. Wikipedia

Delos D. Wickens (1909-1988). Professor of Psychology at Ohio State University; he developed the release-from-proactive-inhibition paradigm and used it to probe how words are semantically encoded. Wikipedia

Frequently Asked Questions

What is proactive inhibition?
Proactive inhibition is the impairment of memory for material learned later by material learned earlier; a person who has studied several lists retains a new list more poorly than a person who studied only the new one, because the earlier learning interferes with the later (Underwood, 1957).

How is proactive inhibition different from retroactive inhibition?
The prefix names the source of the interference. In proactive inhibition, earlier learning reaches forward to disrupt later learning; in retroactive inhibition, later learning reaches back to disrupt earlier learning. Both are forms of interference and typically operate together (Underwood, 1957).

Is proactive inhibition the same as proactive interference?
Yes. Proactive inhibition is the older, verbal-learning term and proactive interference the term preferred in the working-memory literature; they denote the same effect, and Medical Subject Headings indexes it under proactive inhibition (Postman & Underwood, 1973).

What is release from proactive inhibition?
It is the sharp rebound of recall that occurs when the material on a trial shifts to a new encoding class after several trials of one class. Wickens used the size of the release to measure the dimensions along which words are encoded (Wickens, 1970).

Does proactive inhibition explain forgetting in short-term memory?
Largely, yes. Keppel and Underwood showed that the rapid forgetting in short-term retention tasks is small on the first trial and grows as earlier items accumulate, so much of what looked like decay is proactive inhibition (Keppel & Underwood, 1962).

What are the two factors in the two-factor theory?
The two factors are unlearning, the weakening of earlier associations when new ones are learned to the same cue, and response competition, the rivalry between old and new responses at the moment of recall (Postman & Underwood, 1973).

Why are some people more susceptible to proactive inhibition than others?
Susceptibility tracks working-memory capacity and attentional control. People with lower capacity accumulate more proactive interference and are more impaired under divided attention, indicating that resisting interference is an act of executive control (Kane & Engle, 2000).

Where in the brain is proactive interference resolved?
Resolving proactive interference in working-memory recognition recruits the left inferior frontal gyrus of the prefrontal cortex, a region involved in selecting among competing representations (Jonides & Nee, 2006).

References