Intention

Abstract

Intention is the mental state that commits an agent to a course of action and gives that action its purpose. This article treats intention on two fronts cognitive psychology has largely kept apart: the neuroscience of the will, which asks how and when a voluntary action is initiated in the brain, and the psychology of goal pursuit, which asks why a formed intention so often fails to produce the behavior it specifies. It follows the hierarchy of distal, proximal, and motor intentions, the readiness potential and the timing of conscious will, the sense of agency measured through intentional binding, the intention-behavior gap, and the implementation intentions that narrow it. Three demonstrations exercise the Libet timeline, the binding of an action to its effect, and the lift an if-then plan gives to follow-through.

Keywords: intention, volition, readiness potential, goal pursuit

An intention is what stands between wanting something and doing it. A person can desire an outcome without intending to pursue it, and can perform a movement without intending it at all, as in a reflex or a slip; intention is the state that settles on a course of action, sustains it against distraction, and marks the result as something the agent did rather than something that merely happened. Cognitive psychology approaches this state from two directions that rarely meet. One tradition, rooted in electrophysiology, asks what happens in the brain when a person decides to move and when in that sequence the conscious intention appears. The other, rooted in social and health psychology, takes the intention as given and asks why so many sincerely held intentions never translate into action. The sections below build the construct from its levels, follow the will into the brain through the readiness potential and the sense of agency, and then turn to the gap between intention and behavior and the planning that closes it.

What Intention Is

Intention is the mental state in which an agent commits to a course of action, representing a goal together with a settled plan to bring it about and linking the two so that the goal actually guides behavior. It is distinguished from a mere desire, which can be held without any commitment to act, and from a prediction, which forecasts an outcome without owning it; to intend is to have decided, and to hold oneself responsible for the doing. Philosophers of action and cognitive scientists converge on the view that intention plays a functional role no other state fills, namely to initiate action, to sustain it until the goal is met, and to coordinate it with the agent's other plans, which is why an intention persists across the delay between forming a plan and executing it (Pacherie, 2008).

Because a single act of will spans a long arc, from a plan formed days in advance to the muscle commands that execute it in milliseconds, intention is best understood as a hierarchy rather than a single state. Pacherie proposed a dynamic model with three tiers that operate on different timescales and draw on different kinds of representation, as Figure 1 sets out: a distal intention that specifies a goal in abstract terms and is subject to rational deliberation, a proximal intention that anchors that goal to the present situation and triggers action, and a motor intention that specifies the movement itself in a fine-grained sensorimotor code (Pacherie, 2008). The tiers are coupled: a distal intention to travel becomes a proximal intention to leave now, which becomes the motor intention that reaches for the keys.

Types of Intention

The distinction between intentions formed in advance and intentions present only in the acting itself organizes much of the field. A distal or prior intention is one an agent forms and holds before acting, such as a plan made in the morning to phone a colleague after lunch; it can be deliberated, revised, and abandoned while the action it specifies remains in the future. A proximal or immediate intention is the one active at the moment of action, the intention-in-action that actually launches and steers the movement (Pacherie, 2008). The two can come apart: a prior intention may never issue in a proximal one if the moment passes unnoticed, and a proximal intention can arise with no prior plan, as when a person swats at an insect without having decided to.

A second cut concerns what the will actually decides. Brass and Haggard argued that a voluntary action is not a single choice but the resolution of three partly independent components, which they labelled the what, the when, and the whether. The what component selects which action to perform, the when component fixes its timing, and the whether component decides whether to carry the action out at all or to withhold it at the last moment (Brass & Haggard, 2008). The whether component, the capacity to veto an action already prepared, is of particular interest because it locates freedom less in the initiation of action than in its inhibition, a possibility raised by the timing experiments considered next.

The Readiness Potential

The neuroscience of intention began with a slow electrical signal that precedes voluntary movement. Kornhuber and Deecke recorded the electroencephalogram while participants made self-paced movements and, by averaging backward from the moment of muscle activation, uncovered a gradually rising negativity that started more than a second before the movement itself. They named it the Bereitschaftspotential, or readiness potential, and identified it as the brain's preparation for a self-initiated act, the first physiological signature of the will (Kornhuber & Deecke, 1965). The readiness potential is generated in the supplementary motor area and builds up well before any outward sign of movement.

Libet turned this preparatory signal into a claim about consciousness. He asked participants to make a spontaneous flick of the wrist whenever they felt the urge, and to report the position of a revolving spot at the instant they first became aware of the intention to move, a measure now called the W judgment. The readiness potential began around 550 milliseconds before the movement, but the reported moment of conscious intention came only about 200 milliseconds before it, so the brain's preparation preceded the conscious decision by roughly 350 milliseconds (Libet, Gleason, Wright, & Pearl, 1983). Libet took this to show that voluntary acts are initiated unconsciously, while preserving a role for conscious control in the power to veto an action in the final moment before it executes. The demonstration below reconstructs this timeline and lets the key latencies be varied.

Later work pushed the antecedents of choice earlier still and then challenged their interpretation. Using functional magnetic resonance imaging, Soon and colleagues found that patterns in prefrontal and parietal cortex predicted which of two buttons a person would press up to several seconds before the person reported deciding, though the prediction was well above chance rather than perfect (Soon, Brass, Heinze, & Haynes, 2008). Schurger and colleagues then offered a reinterpretation that dissolves much of the paradox. On their account the readiness potential is not the trace of a decision already made but the rising limb of a stochastic accumulator: spontaneous neural activity fluctuates, a self-initiated movement is triggered when that fluctuation happens to cross a threshold, and averaging backward from the movement necessarily reveals a gradual rise that was, until threshold, mostly noise (Schurger, Sitt, & Dehaene, 2012). The readiness potential, on this view, marks not when the brain decided but when accumulating activity tipped past a bound.

Intentional Binding and the Sense of Agency

Alongside the timing of intention sits the feeling that accompanies it: the sense of agency, the experience that one is the author of one's actions and their effects. Because this feeling is difficult to interrogate directly, Haggard and colleagues devised an implicit measure. They found that when a voluntary action causes an effect, such as a keypress that triggers a tone, the perceived time of the action shifts later and the perceived time of the effect shifts earlier, so the two are drawn together in subjective time. This intentional binding did not occur when the movement was induced involuntarily by transcranial magnetic stimulation, where the perceived events instead repelled one another (Haggard, Clark, & Kalogeras, 2002). The compression of perceived time is thus a marker of voluntary causation, present when the agent intends the action and absent when the same movement is imposed from outside.

Intentional binding has become a workhorse measure of agency precisely because it does not ask people how in control they feel, a question their answers to which are easily biased, but reads agency off a subtle distortion of perceived timing. The demonstration below contrasts a voluntary keypress with an involuntary movement and shows how the perceived interval between action and tone compresses in the first case and expands in the second.

The Intention-Behavior Gap

The second research tradition takes the intention as settled and asks whether it produces the behavior it specifies. The answer is a sobering one for any account that treats intention as the direct cause of action. Ajzen's theory of planned behavior makes intention the proximal determinant of behavior, itself shaped by attitudes toward the behavior, subjective norms, and perceived behavioral control, so that a stronger intention should mean a higher probability of acting (Ajzen, 1991). Yet intention explains far less of behavior than the theory's centrality to it would suggest. Sheeran's review of the evidence found that intention accounts for only about a quarter of the variance in later behavior, and that the discrepancy is asymmetric: people who form an intention frequently fail to act, whereas people who do not intend to act rarely do (Sheeran, 2002).

The gap is not merely correlational slack but a causal shortfall, as experiments that change intention reveal. Webb and Sheeran pooled studies that manipulated intention experimentally and found that a medium-to-large change in intention produced only a small change in behavior, direct evidence that intending more does not translate proportionally into doing more (Webb & Sheeran, 2006). Sheeran and Webb later mapped where intentions break down, distinguishing people who fail to get started from those who get started but fall short, and locating the gap in the phase between forming a goal and enacting it rather than in the sincerity of the goal itself (Sheeran & Webb, 2016). The problem, in short, is not motivation but the translation of motivation into action.

Implementation Intentions

The most effective known remedy for the gap is a change in the form of the intention rather than its strength. Gollwitzer distinguished a goal intention, which specifies an outcome in the form I intend to reach Z, from an implementation intention, which specifies in advance the situation in which one will act and the response one will make, in the form if situation X arises, then I will perform response Y. By linking a concrete cue to a concrete action, an implementation intention delegates the initiation of behavior to the environment, so that the cue elicits the action automatically without a further act of deliberation (Gollwitzer, 1999). The plan does not make the person want the goal more; it arranges for the wanting to be acted on when the moment comes.

The size of the effect is well established. Gollwitzer and Sheeran's meta-analysis of ninety-four studies found that forming an implementation intention had a medium-to-large effect on goal attainment, of about d equals 0.65, over and above the effect of the goal intention alone (Gollwitzer & Sheeran, 2006). Because the plan works by automating initiation, its benefit is largest for goals people mean to pursue but reliably forget or postpone, and smaller where the behavior was going to happen anyway. The demonstration below models this lift, letting a baseline follow-through rate be set and showing how an if-then plan raises the expected probability of acting.

Prospective Memory

Between forming an intention and acting on it lies a problem of memory: an intention to act later must be retained and then retrieved at the right moment, often while attention is occupied by something else. Prospective memory is the capacity to remember to carry out a delayed intention, and its failures, forgetting to pass on a message or to take a medication, are among the most common lapses of everyday cognition. McDaniel and Einstein proposed a multiprocess framework in which the retrieval of a delayed intention can proceed in two ways. When a distinctive cue coincides with a focal task, retrieval can be relatively automatic, the cue spontaneously bringing the intention to mind; when the cue is not focal or is poorly specified, retrieval instead demands effortful strategic monitoring that draws on limited attentional resources (McDaniel & Einstein, 2000). The framework explains why some delayed intentions pop into mind effortlessly while others require the person to keep checking, and it connects the goal-pursuit literature to the mechanics of memory, since an implementation intention works in part by making the triggering cue focal enough to support automatic retrieval.

The Neural Basis of Volition

Converging evidence locates the initiation of voluntary action in a medial frontal network, above all the supplementary motor area and pre-supplementary motor area that generate the readiness potential, together with parietal regions that appear to underwrite the conscious experience of intending. Direct electrical stimulation makes the dissociation vivid. Desmurget and colleagues found that stimulating the inferior parietal cortex in awake neurosurgical patients produced a conscious urge to move a body part, and at higher intensities the belief that the movement had occurred, though no muscle activity was recorded, whereas stimulating the premotor cortex produced actual movements that patients denied having made (Desmurget et al., 2009). Intention to move and the movement itself can therefore be pulled apart by where the brain is stimulated, with parietal cortex supplying the felt intention and premotor cortex the motor output.

Recordings at the level of single neurons close the loop between the readiness potential and the moment of awareness. Fried and colleagues recorded from neurons in the medial frontal cortex of patients performing a Libet-style task and found cells whose firing rate changed progressively before the reported intention to move, with enough information in a few hundred such neurons to predict the impending action, and its approximate timing, before the person reported deciding (Fried, Mukamel, & Kreiman, 2011). Haggard has synthesized this body of work into an account of volition as a set of neurocognitive functions rather than a single faculty, in which distinct processes generate, time, and gate action, and in which the sense of a unified will is a late construction rather than the cause of the movement (Haggard, 2008; Haggard, 2019). Table 1 sets out the three components of that account and the questions each resolves.

Note. The three components are partly independent, so a voluntary action is the joint resolution of what, when, and whether rather than a single indivisible choice. Adapted from Brass and Haggard (2008) and Haggard (2008).

Worked Example

The effect of an implementation intention can be made precise with a simple threshold model that the demonstration above implements. Suppose a person's tendency to follow through on a goal is a continuous latent quantity, and that they act whenever that quantity exceeds a fixed threshold. A baseline follow-through rate then corresponds to a point on the standard normal curve: if 40 percent of people with a given goal act on it, the baseline probability is P0 equals 0.40, which maps to a latent value of z0 equals the inverse normal of 0.40, about minus 0.25. Forming an implementation intention shifts the latent tendency upward by a fixed amount equal to the meta-analytic effect size, d equals 0.65.

Adding that shift gives a new latent value of z1 equals minus 0.25 plus 0.65, which is about 0.40, and converting back through the normal curve gives P1 equals the normal probability at 0.40, or about 0.65. Follow-through therefore rises from 40 percent to about 65 percent, a gain of roughly 25 percentage points, from a plan that adds no new motivation and only specifies when and where to act (Gollwitzer & Sheeran, 2006). The S-shaped geometry of the normal curve makes the gain largest for goals that start near the middle of the range, where the curve is steepest, and smaller for goals that were already very likely or very unlikely to be enacted. The demonstration above lets the baseline rate be varied so this pattern can be traced across the full range.

Discussion

The two traditions this article has traced rarely cite one another, yet they describe the same arc from the opposite ends. The neuroscience of the will works forward from the brain, showing that a voluntary movement is prepared by activity that begins before the agent is aware of intending it (Libet et al., 1983), that this preparatory signal may be the crossing of a threshold by spontaneous fluctuation rather than the readout of a prior decision (Schurger et al., 2012), and that the felt authorship of action is a reconstruction indexed by subtle distortions of perceived time (Haggard et al., 2002). The psychology of goal pursuit works backward from behavior, showing that a consciously formed intention is a weak and asymmetric predictor of what a person will actually do (Sheeran, 2002), and that the shortfall lies in the translation of a goal into action, a phase that a well-formed if-then plan can automate (Gollwitzer, 1999). Read together they converge on a deflationary but productive picture: intention is neither the uncaused first mover of folk psychology nor an illusion, but a real and functionally specific set of processes that prepare, time, gate, and monitor action, and whose known failure modes, the unnoticed initiation and the unenacted plan, are open to measurement and, in the second case, to intervention. The practical payoff is concentrated in that second case, where understanding the intention-behavior gap as a problem of translation rather than of will has yielded one of the more robust behavior-change techniques in psychology.

Glossary

Accumulator model.

Schurger's account in which a self-initiated movement is triggered when spontaneously fluctuating neural activity crosses a threshold, so the readiness potential reflects that crossing rather than a prior decision.

Bereitschaftspotential.

The readiness potential; a slow negative electrical potential over the supplementary motor area that builds up before a self-initiated movement.

Distal intention.

A prior intention that specifies a goal in abstract terms and is held before action, open to deliberation and revision while the action remains in the future.

Goal intention.

An intention that specifies a desired outcome, in the form I intend to reach a stated goal, without specifying when or where the action will be taken.

Implementation intention.

A plan of the form if situation X arises then I will perform response Y, which links a cue to an action so that initiation becomes automatic.

Intention-behavior gap.

The discrepancy between the intentions people form and the behavior they enact, in which a large share of good intentions fail to translate into action.

Intention.

A mental state that commits an agent to a course of action, representing a goal together with a settled plan that guides and sustains the behavior toward it.

Intentional binding.

The compression of perceived time between a voluntary action and its effect, used as an implicit measure of the sense of agency.

Motor intention.

The lowest tier of the intention hierarchy, a fine-grained sensorimotor specification that guides the movement as it is executed.

Prospective memory.

The capacity to remember to carry out a delayed intention at the appropriate later moment, often while attention is engaged elsewhere.

Proximal intention.

The intention active at the moment of action, which anchors a goal to the present situation and initiates and steers the movement.

Readiness potential.

The gradually rising brain potential that precedes a voluntary movement by up to a second or more; the physiological signature of motor preparation.

Sense of agency.

The experience of being the author of one's own actions and their effects, measured implicitly through phenomena such as intentional binding.

Theory of planned behavior.

Ajzen's framework in which intention is the proximal determinant of behavior, shaped by attitudes, subjective norms, and perceived behavioral control.

Veto.

The capacity to withhold an action that has already been prepared, associated with the whether component of intentional action and sometimes called free won't.

Volition.

The set of processes by which action is generated from within rather than driven by an external stimulus; the will, treated as a family of functions rather than a single faculty.

W judgment.

In Libet's paradigm, a participant's report of the clock position at the moment they first became aware of the intention to move.

What, when, whether model.

Brass and Haggard's decomposition of a voluntary action into partly independent decisions about which action to perform, when to perform it, and whether to perform it at all.

Key Researchers

Icek Ajzen. Developed the theory of planned behavior, the framework that makes intention the proximal predictor of behavior; Professor Emeritus in the Department of Psychological and Brain Sciences at the University of Massachusetts Amherst, ORCID 0000-0001-8101-6300. ORCID · Google Scholar · Faculty page · Wikipedia

Peter M. Gollwitzer. Originated the theory of implementation intentions and, with Sheeran, established their effect on goal attainment by meta-analysis; Professor of Psychology at New York University and the University of Konstanz, ORCID 0000-0003-4872-2929. ORCID · Google Scholar · Faculty page · Wikipedia

Patrick Haggard. Developed the intentional-binding measure of the sense of agency and the what, when, whether model of volition, and has synthesized the neuroscience of human will; Professor of Cognitive Neuroscience at University College London, ORCID 0000-0001-7798-793X. ORCID · Google Scholar · Faculty page · Wikipedia

Benjamin Libet (1916-2007). Devised the timing experiment that placed the readiness potential before the reported moment of conscious intention, reframing the debate over conscious will; worked in the Department of Physiology at the University of California, San Francisco. Wikipedia

Elisabeth Pacherie. Formulated the dynamic hierarchical model of intention that distinguishes distal, proximal, and motor tiers of the will; Directrice de Recherche at the CNRS, based at the Institut Jean Nicod in Paris, ORCID 0000-0003-2337-581X. ORCID · Google Scholar · Faculty page · Wikipedia

Aaron Schurger. Proposed the stochastic-accumulator reinterpretation of the readiness potential, reframing self-initiated movement as threshold-crossing by spontaneous activity; on the faculty of Chapman University and formerly of the INSERM Cognitive Neuroimaging Unit in Paris, ORCID 0000-0003-2985-3253. ORCID · Google Scholar · Faculty page

Paschal Sheeran. Charted the intention-behavior gap through conceptual review and meta-analysis and, with colleagues, the interventions that narrow it; Stephen J. Walsh Distinguished Professor in the Department of Psychology and Neuroscience at the University of North Carolina at Chapel Hill, ORCID 0000-0001-7449-4590. ORCID · Google Scholar · Faculty page

Frequently Asked Questions

What is intention in cognitive psychology?
Intention is the mental state that commits an agent to a course of action, representing a goal together with a settled plan that guides and sustains the behavior toward it. It is distinguished from a mere desire, which carries no commitment to act, and from a prediction, which forecasts an outcome without owning it (Pacherie, 2008).

What is the readiness potential?
The readiness potential, or Bereitschaftspotential, is a slow rising brain potential over the supplementary motor area that begins up to a second or more before a self-initiated movement. It was discovered by Kornhuber and Deecke and is regarded as the physiological signature of motor preparation (Kornhuber & Deecke, 1965).

Did Libet's experiment disprove free will?
No; it showed only that the readiness potential begins before the reported moment of conscious intention, which reframed the debate rather than settling it. A later account reinterprets the signal as accumulating spontaneous activity crossing a threshold, not as an unconscious decision made in advance (Schurger, Sitt, & Dehaene, 2012).

What is intentional binding?
Intentional binding is the compression of perceived time between a voluntary action and its effect, so that the action seems later and the effect earlier than they truly were. Because this compression does not occur for involuntary movements, it serves as an implicit measure of the sense of agency (Haggard, Clark, & Kalogeras, 2002).

What is the intention-behavior gap?
The intention-behavior gap is the discrepancy between the intentions people form and the behavior they carry out. Intention accounts for only about a quarter of the variance in later behavior, and the gap is asymmetric, since people who intend to act often fail while people who do not intend to act rarely do (Sheeran, 2002).

How do implementation intentions work?
An implementation intention specifies in advance a situation and a response, in the form if situation X arises then I will perform response Y, which links a cue to an action so that initiation becomes automatic. A meta-analysis found a medium-to-large effect on goal attainment beyond that of the goal intention alone (Gollwitzer & Sheeran, 2006).

What is the difference between a distal and a proximal intention?
A distal intention is a prior intention that fixes an abstract goal before action and can be deliberated and revised, whereas a proximal intention is the intention active at the moment of action that initiates and steers the movement. The two can come apart, as when a prior plan is never enacted or an action is taken with no prior plan (Pacherie, 2008).

What is prospective memory?
Prospective memory is the capacity to remember to carry out a delayed intention at the right later moment. Its retrieval can be relatively automatic when a distinctive cue coincides with the ongoing task, or effortful and monitoring-dependent when the cue is not focal (McDaniel & Einstein, 2000).

References