Affect

Abstract

Affect, which MeSH classifies under emotions, is in contemporary psychology the more basic notion: the raw felt quality of experience, the pleasant or unpleasant, activated or quiet tone that colors every mental state and out of which specific emotions and moods are built. This article surveys the circumplex and core-affect models that describe affect along dimensions of valence and arousal; the argument that affective reactions can precede and occur independently of cognition; the separability of positive and negative affect; the affect-as-information and affect-infusion accounts of how mood shapes judgment; the broaden-and-build theory of positive affect; and the measurement of affect and its temporal dynamics. Three interactive demonstrations let the reader explore the two-axis structure of affect, mood misattribution in judgment, and the dynamics of affect over time.

Keywords: affect, core affect, valence and arousal, mood, affect-as-information

Affect is the most general term in the psychology of feeling: it names the felt, valenced quality of a mental state, the sense that things are going well or badly and that the organism is stirred up or settled. Where an emotion is a brief, object-directed episode and a mood is a diffuse, longer-lasting background state, affect is the property they share — the elementary feeling-tone that both involve. A growing body of theory treats this feeling-tone as a psychological primitive: a basic, irreducible ingredient of the mind that is always present and that more complex emotional states are constructed from, rather than a byproduct of them (Barrett & Bliss-Moreau, 2009). On this view the always-on affective state, sometimes called core affect, is a continuously varying neurophysiological condition describable by how pleasant it is and how activated it is, and an emotion is what results when that core affect is attributed to an object and categorized with a learned concept (Russell & Barrett, 1999; Russell, 2003). This inverts the folk picture: rather than emotions producing feelings, a primitive feeling is the material from which emotions and moods are made (Barrett, 2017).

Types of Affect

The Medical Subject Headings vocabulary files Affect (D000339) beneath Emotions in tree F01.470.047 and lists one narrower descriptor directly under it, shown in Table 1. The single subtype reflects indexing practice, not a scientific taxonomy of affect: MeSH is a controlled vocabulary for retrieving literature, so its placement of affect under emotions is an indexing convention that runs opposite to the contemporary theoretical view, in which affect is the superordinate substrate and specific emotions are the narrower kinds. The entry below is glossed from its ordinary psychological meaning; it does not yet have its own article on this site.

Affective Primacy

A foundational question is whether affect depends on prior cognition or can arise without it. Robert Zajonc argued for affective primacy: that affective reactions are often the first response to a stimulus, are made quickly and with confidence, and can occur before and independently of the cognitive processes that identify what the stimulus is — his slogan that preferences need no inferences (Zajonc, 1980). His evidence included the mere-exposure effect, in which repeated exposure to a stimulus increases liking for it even when observers cannot recognize having seen it, so that a preference forms without any conscious cognition of the object. The claim was deliberately provocative against the then-dominant view that emotion is always post-cognitive, a product of appraisal, and it set up a debate about the temporal and causal relation between affect and cognition that appraisal theorists such as Lazarus contested directly.

The modern resolution treats the two as deeply intertwined rather than strictly ordered. That affect can be fast and pre-recognition is well supported, but affect and cognition share neural substrates and continuously shape each other, which is why affect is now studied less as a stage that comes before or after thought than as a pervasive influence woven through perception, attention, memory, and judgment. This integrative stance — that affect is not a separable add-on but a constituent of most mental processes — has become a defining commitment of contemporary affective science (Dukes et al., 2021).

The Structure of Affect

If affect is describable at all, what are its dimensions? The dominant answer places felt states in a plane. James Russell's circumplex model proposed that the quality of any affective state is captured by two dimensions — valence, from unpleasant to pleasant, and arousal, from calm to activated — so that affect words arrange themselves around a circle: excitement at high-arousal positive, contentment at low-arousal positive, distress at high-arousal negative, and depression at low-arousal negative (Russell, 1980). Working independently, David Watson and Auke Tellegen proposed a consensual structure of mood in which the same space is often described by a rotated pair of axes, positive affect and negative affect, and showed that self-reported mood terms recover this two-dimensional structure across many datasets (Watson & Tellegen, 1985). The circumplex and the two-factor descriptions are rotations of one plane, not rival geometries, and the circumplex has since been given an explicit affective-neuroscience reading in which valence and arousal correspond to interacting neural systems (Posner et al., 2005).

The rotation matters, however, because it embodies a substantive claim about whether positivity and negativity are one axis or two. John Cacioppo and Gary Berntson's evaluative space model argues that positive and negative affect rest on partly separable substrates that can be activated independently, so that the bipolar assumption — that more positive necessarily means less negative — fails: a person can feel both, neither, or one without the other (Cacioppo & Berntson, 1994). Ambivalence (high on both) and disengagement (low on both) share a net valence near zero yet are psychologically opposite states, which a single bipolar scale cannot distinguish. The first demonstration lets a reader set positive and negative affect independently and see why one number is not enough.

Positive and negative affect vary separately, so the state is a point in a plane rather than on a line. A single bipolar scale reports only the difference PA−NA, and so cannot tell high-on-both ambivalence from low-on-both disengagement — two states with the same net valence but opposite activation. The separability of the two substrates is the core of the evaluative space model (Cacioppo & Berntson, 1994) and the two-factor structure of mood (Watson & Tellegen, 1985).

Affect in Cognition and Judgment

Because affect is nearly always present, one of its most studied roles is its influence on judgments that seem unrelated to feeling. The affect-as-information account of Norbert Schwarz and Gerald Clore holds that people often evaluate a target by implicitly asking “How do I feel about it?” and reading their current affective state as an answer, so that an incidental mood can infuse an unrelated judgment (Schwarz & Clore, 1983). Their classic result is a misattribution study: respondents reported higher life satisfaction on sunny days than rainy ones, but the weather effect vanished when the interviewer first drew attention to the weather, because respondents then discounted their mood as irrelevant to the life-satisfaction question. The feeling was the same; what changed was whether it was treated as information about the target. The second demonstration reproduces this logic, letting a reader vary a mood and its attributed cause and watch the judgment infuse or snap back to neutral.

With no salient cause, an unrelated pleasant or unpleasant mood infuses an unrelated judgment — rate life satisfaction and the answer tracks how you happen to feel. Attribute that same mood to a transient, irrelevant cause such as the weather, and its informational value is discounted: the judgment snaps back toward the neutral anchor even though the feeling is unchanged. This misattribution result is the signature of the affect-as-information account (Schwarz & Clore, 1983).

Joseph Forgas embedded this and related findings in a broader theory of when mood matters. His affect infusion model specifies that affect infuses cognition most when a judgment requires open, constructive, substantive processing — elaborating, imagining, inferring — and least when it can be made by direct retrieval of a stored evaluation or by a motivated, tightly directed strategy (Forgas, 1995). The model thus predicts the paradox that more thinking can mean more mood influence, not less, because the constructive processing that thorough judgment demands is exactly what lets current affect seep in. A related demonstration of affect's reach into cognition is affective forecasting: people predicting how they will feel about future outcomes systematically overestimate the intensity and duration of their reactions, a bias that shapes decisions made in the present (Wilson & Gilbert, 2005).

Positive Affect and Its Functions

Much of the older literature treated affect through the lens of negative states and stress, but positive affect turns out to have distinctive functions of its own. Barbara Fredrickson's broaden-and-build theory holds that positive emotions widen the momentary thought-action repertoire — joy invites play, interest invites exploration, contentment invites savoring and integration — and that these broadened states, accumulated over time, build enduring personal resources: physical, intellectual, social, and psychological (Fredrickson, 2001). The theory reframes positive affect not as a mere feel-good byproduct but as an evolved mechanism for growth, and it contrasts sharply with the narrowing, action-focused character of negative affect.

Positive and negative affect also anchor the study of subjective well-being, the scientific treatment of how people evaluate their lives. Ed Diener's foundational synthesis separated well-being into a cognitive component, judgments of life satisfaction, and an affective component, the balance of positive over negative affect experienced across time, establishing that a good life as people actually assess it has an irreducibly affective ingredient (Diener, 1984). This partition mirrors the structural point above: because positive and negative affect are separable, well-being is not captured by a single pleasant-minus-unpleasant score.

Measuring Affect and Its Dynamics

Affect is measured both by self-report and by tools designed to bypass language. The most widely used self-report instrument, the Positive and Negative Affect Schedule (PANAS) of Watson, Clark, and Tellegen, asks respondents to rate a set of affect adjectives and yields separate positive-affect and negative-affect scores, operationalizing the two-factor structure directly (Watson et al., 1988). To capture the valence-and-arousal dimensions without words, Margaret Bradley and Peter Lang developed the Self-Assessment Manikin, a pictorial scale on which a figure ranges from frowning to smiling for valence and from calm to excited for arousal, allowing affect to be rated quickly and across cultures and ages (Bradley & Lang, 1994).

Averages, however, miss much of what matters about affect, because affect unfolds in time. The study of affect dynamics characterizes a person not only by their mean level of positive and negative affect but by how that affect moves: its variability (the size of the swings), its inertia (how strongly each state carries into the next), and its rate of return to baseline. Peter Kuppens and Philippe Verduyn review evidence that these dynamic features carry information beyond the mean — high variability and high inertia, in particular, are associated with lower well-being and with vulnerability to psychopathology (Kuppens & Verduyn, 2017). The final demonstration generates an affect trajectory from a simple dynamic model, letting a reader vary its variability and inertia and watch the shape of the feeling change.

observed swing (mean abs. deviation) = 0.21

Two people can share an average mood yet differ in how it moves. Variability sets the size of the swings; inertia sets how strongly each state carries into the next, smoothing the line and slowing its return to baseline. High variability and high inertia — large, sticky departures from baseline — are the affect-dynamic signatures linked to lower well-being and psychopathology (Kuppens & Verduyn, 2017).

Worked Example

The affect-as-information account makes a quantitative prediction that can be checked with arithmetic: how far an incidental mood shifts an unrelated judgment depends on the weight the judge gives that mood as information. Model a reported evaluation as a neutral anchor plus a fraction of the felt mood's departure from neutral. Let the judgment be J = A + w(M − A), where A = 50 is the neutral anchor on a 0–100 scale, M is the current mood on the same scale, and w is the informational weight — how far the mood is treated as relevant to the target. When the mood is fully treated as information, w = 1 and the judgment equals the mood; when the mood is discounted because it is attributed to an irrelevant cause, w falls toward 0 and the judgment returns to the anchor.

Take a pleasant mood of M = 78. With no salient alternative cause the mood is read as information about the target, w = 1.0, and J = 50 + 1.0(78 − 50) = 50 + 28 = 78: the judgment simply tracks the feeling. Now let the judge's attention be drawn to the weather, an irrelevant transient cause, so that most of the mood's informational value is discounted, w = 0.15. Then J = 50 + 0.15(78 − 50) = 50 + 0.15 × 28 = 50 + 4.2 = 54.2, which rounds to 54. The very same feeling now moves the judgment by only 4 points instead of 28. A weaker attribution to the survey room, w = 0.2, gives J = 50 + 0.2 × 28 = 55.6, rounding to 56.

The example makes the theory's claim precise. The mood itself never changed; what changed was w, the degree to which the feeling was treated as data about the target. This is why misattribution manipulations abolish mood effects on judgment without touching the mood, and why the same incidental affect can dominate one evaluation and barely touch another depending on whether the judge reads it as relevant (Schwarz & Clore, 1983).

Discussion

Across these literatures a coherent picture of affect emerges even where specific theories disagree. Affect is best understood as a basic, continuously present, low-dimensional property of mind — well described by valence and arousal — that both precedes and pervades cognition rather than merely following it (Zajonc, 1980; Russell, 2003; Barrett & Bliss-Moreau, 2009). Its positive and negative aspects are partly independent, so affect is not a single scale from misery to bliss but at least a plane, with consequences for how well-being is measured and how mixed states such as ambivalence are understood (Cacioppo & Berntson, 1994; Diener, 1984).

The functional story is equally two-sided. Affect is not noise to be filtered out of rational judgment; it is information the mind routinely uses, sometimes appropriately and sometimes not, as the misattribution and affective-forecasting findings show (Schwarz & Clore, 1983; Forgas, 1995; Wilson & Gilbert, 2005). Positive affect in particular does constructive work, broadening cognition and building resources rather than simply signaling that all is well (Fredrickson, 2001). The recurring theme is that affect is a substrate and a signal at once: the raw material from which emotions are built and a continuously consulted source of information about the world and the self.

Current Directions

Three currents mark the contemporary study of affect. The first is the rise of affectivism — a programmatic argument that affect and emotion are not a specialized subfield but a pervasive determinant of perception, cognition, decision, and social behavior, and that psychology and neuroscience advance faster when affect is treated as central rather than peripheral (Dukes et al., 2021). This reframes many of the judgment and cognition findings above as instances of a general principle rather than isolated curiosities.

The second is methodological and taxonomic. Data-driven studies of naturalistic emotional responses recover a high-dimensional, gradient-structured space of dozens of reported states, pressing the question of how the compact valence-arousal description of affect relates to the far richer catalog of distinctly labeled feelings people report (Cowen & Keltner, 2017). In parallel, multivariate decoding of brain activity has moved past asking whether single regions signal single feelings, showing instead that distributed activity patterns can be decoded into affective categories — evidence that affect is represented at the level of networks rather than dedicated centers (Kragel & LaBar, 2016). The third, running through both, is dynamic: the shift from characterizing a person's affect by its average to characterizing it by its movement over time, a change of unit that has already proven informative for well-being and psychopathology (Kuppens & Verduyn, 2017).

Common Misconceptions

Affect and emotion are the same thing.

Affect is the broad, always-present feeling-tone; an emotion is a brief, object-directed episode built from it, and a mood is a diffuse, objectless affective state. Many accounts treat core affect as the primitive from which emotions are constructed (Russell & Barrett, 1999; Barrett & Bliss-Moreau, 2009).

Positive and negative affect are opposite ends of one scale.

They rest on partly separable systems, so a person can be high on both (ambivalent) or low on both (disengaged) — states a single bipolar scale cannot tell apart (Cacioppo & Berntson, 1994).

Feelings only follow thinking.

Affective reactions can be fast and can precede recognition of what a stimulus is, as the mere-exposure effect shows, in Zajonc's phrase that preferences need no inferences (Zajonc, 1980).

Affect is just interference that good judgment removes.

Affect is routinely used as information, and constructive, effortful judgment can admit more mood influence, not less; positive affect also does constructive cognitive work (Forgas, 1995; Fredrickson, 2001).

Glossary

Affect infusion model.

Forgas's account specifying that mood infuses judgment most when the task calls for open, constructive processing and least under direct retrieval or motivated strategies.

Affect-as-information.

The view that people read their current affective state as data about a target (“How do I feel about it?”), so incidental mood can infuse unrelated judgments.

Affect.

The basic felt, valenced quality of a mental state; the general feeling-tone shared by emotions and moods.

Affective forecasting.

Predicting one's future affective reactions; forecasts systematically overestimate the intensity and duration of feelings.

Affective primacy.

Zajonc's claim that affective reactions can be generated quickly and independently of, and prior to, the cognition that identifies a stimulus.

Arousal.

The activation dimension of affect, running from calm to highly activated; one of the two axes of the circumplex.

Broaden-and-build theory.

Fredrickson's theory that positive affect widens the momentary thought-action repertoire and, over time, builds enduring personal resources.

Circumplex model.

Russell's representation of affect as a circle in a plane of valence and arousal, with similar states placed as near neighbors.

Core affect.

A continuously present, primitive neurophysiological state of valence and arousal, the substrate from which emotions and moods are built.

Emotional inertia.

The degree to which an affective state carries over into the next moment; higher inertia means slower return to baseline and is linked to lower well-being.

Evaluative space model.

Cacioppo and Berntson's proposal that positive and negative affect rest on partly independent substrates rather than a single bipolar dimension.

Mood.

A diffuse, longer-lasting affective state that lacks a clear object, distinguished from the brief, object-directed episode of an emotion.

Negative affect (NA).

The dimension of aversive, distressing engagement (e.g., distress, nervousness), measured separately from positive affect on the PANAS.

Positive affect (PA).

The dimension of pleasurable, activated engagement (e.g., enthusiasm, alertness), measured separately from negative affect on the PANAS.

Subjective well-being.

A person's own evaluation of their life, comprising a cognitive component (life satisfaction) and an affective component (the balance of positive over negative affect).

Valence.

The pleasantness dimension of affect, running from unpleasant to pleasant; one of the two axes of the circumplex.

Key Researchers

Lisa Feldman Barrett (contemporary). Psychologist at Northeastern University; she argued that affect is a psychological primitive and built the theory of constructed emotion, in which discrete emotions are categorizations of core affect. ORCID - Google Scholar - Faculty Page - Wikipedia - Wikidata

John T. Cacioppo (1951-2018). Psychologist at the University of Chicago; he co-developed the evaluative space model, arguing that positive and negative affect are separable, partly independent substrates. ORCID - Wikipedia - Wikidata

Ed Diener (1946-2021). Psychologist at the University of Illinois at Urbana-Champaign; a foundational figure in subjective well-being research, separating its affective and cognitive components. Google Scholar - Wikipedia - Wikidata

Joseph P. Forgas (contemporary). Psychologist at the University of New South Wales; he built the affect infusion model, specifying when and how mood infuses judgment and social cognition. ORCID - Google Scholar - Faculty Page

Barbara L. Fredrickson (contemporary). Psychologist at the University of North Carolina at Chapel Hill; she proposed the broaden-and-build theory of positive affect. ORCID - Google Scholar - Faculty Page - Wikipedia - Wikidata

James A. Russell (contemporary). Psychologist at Boston College; he proposed the circumplex model of affect and the theory of core affect. ORCID - Google Scholar - Wikipedia - Wikidata

Norbert Schwarz (contemporary). Psychologist at the University of Southern California; he developed the affect-as-information account of judgment. ORCID - Google Scholar - Faculty Page - Wikipedia - Wikidata

Auke Tellegen (1930-2024). Psychologist at the University of Minnesota; he co-authored the consensual structure of mood and the PANAS. Wikipedia - Wikidata

David Watson (contemporary). Psychologist at the University of Notre Dame; he co-developed the PANAS and the two-dimensional positive-affect/negative-affect model of mood. Google Scholar - Faculty Page - Wikipedia

Robert B. Zajonc (1923-2008). Psychologist at Stanford University and the University of Michigan; he argued for affective primacy, that affect can precede and occur independently of cognition. Wikipedia - Wikidata

Frequently Asked Questions

What is affect in psychology?
Affect is the basic felt, valenced quality of a mental state: the pleasant or unpleasant, activated or calm feeling-tone that colors experience. It is the general category of feeling from which specific emotions and moods are built, and many theories treat it as a psychological primitive (Barrett & Bliss-Moreau, 2009).

How is affect different from emotion and mood?
An emotion is a brief episode directed at a specific object; a mood is a diffuse, longer-lasting state without a clear object; affect is the underlying feeling-tone both share. On core-affect accounts, an emotion is core affect attributed to an object and categorized (Russell & Barrett, 1999; Russell, 2003).

What are the two dimensions of affect?
Most models describe affect with valence (unpleasant to pleasant) and arousal (calm to activated), placing all felt states in a plane, an arrangement known as the circumplex model (Russell, 1980).

Can positive and negative affect occur together?
Yes. They rest on partly separable systems, so a person can be high on both (ambivalent) or low on both (disengaged), which a single bipolar scale cannot capture (Cacioppo & Berntson, 1994).

Does affect come before or after thinking?
Affect can be fast and can precede recognition of a stimulus, as the mere-exposure effect shows, but affect and cognition also continuously shape each other rather than running in a strict order (Zajonc, 1980).

How does mood influence unrelated judgments?
People often read their current mood as information about whatever they are judging, so an incidental feeling infuses the judgment, unless the mood is attributed to an irrelevant cause, in which case its influence disappears (Schwarz & Clore, 1983).

How is affect measured?
By self-report scales such as the PANAS, which yields separate positive- and negative-affect scores (Watson et al., 1988), and by non-verbal pictorial scales such as the Self-Assessment Manikin, which rate valence and arousal directly (Bradley & Lang, 1994).

What is affect dynamics?
Affect dynamics studies how feeling moves over time (its variability, inertia, and return to baseline) rather than only its average level, and finds that these dynamic features predict well-being and psychopathology (Kuppens & Verduyn, 2017).

References