What 'High Agency' Hides: Staying Still Feels Heavier Than Moving

Trait words for drive keep being renamed, from grit to high agency. This paper argues that they describe an outcome, sets out the mechanism the evidence supports, and draws out what it means for how children are raised and taught.

Abstract. "High agency" has become the preferred trait word for drive in founder culture, and it is moving into parenting and education. As with grit before it, the term names an observed pattern of behaviour and is then treated as the cause of that behaviour. This paper reviews four bodies of evidence: research on trait labels, cost-benefit models of effort, research on comparison and self-evaluation, and research on intrinsic motivation. Together they are consistent with a simpler account, which this paper proposes.

A person acts when staying still is perceived as heavier than moving. Push adds weight to staying still: a gap between where one stands and where one needs to stand, measured against a reference. Pull makes moving lighter: curiosity that finds growing competence. Identity sets much of the weight of failure, and energy and environment condition the whole system. We map common trait words onto these parts, state five testable predictions, and draw implications for children, including one specific to a period in which AI makes finished work cheap to produce.

1. A description promoted to a cause

In its popular form, high agency is described as the ability to shape one's circumstances rather than be shaped by them. Popular accounts break it into components such as clear thinking, a bias to action and disagreeability, and add an immigrant mentality and sustained high energy. They rarely test it. A researcher at the University of Melbourne has noted that the idea overlaps with long-studied constructs such as perceived control and self-efficacy, and that these vary with age, income and education.

Psychology has met this problem before. Truman Kelley (1927) named the jangle fallacy: two measures with different names that capture the same thing. Grit, introduced by Duckworth, Peterson, Matthews and Kelly (2007), is the recent case. A meta-analysis by Credé, Tynan and Harms (2017), pooling 88 samples and 66,807 participants, found grit very strongly correlated with conscientiousness, only moderately related to performance and retention, and concluded that interventions designed to raise grit are likely to have weak effects. Rotter's (1966) locus of control and Bandura's (1977) self-efficacy sit in the same family.

A trait label can summarise behaviour accurately and still give no handle for changing it. That is the practical difficulty for parents and teachers. A description of how a driven person appears from outside says nothing about what produces the drive, so it cannot be taught or designed for.

2. The mechanism: which option is heavier

Kurt Lewin (1936) wrote behaviour as a function of the person and the environment, B = f(P, E), and later (1951) described any stable state as a balance between driving and restraining forces. Expectancy-value theory (Eccles et al., 1983) includes cost as a component of every choice: the effort required, what must be given up, and the emotional price of trying. Kurzban, Duckworth, Kable and Myers (2013) argued that the sense of effort is itself an opportunity-cost signal, the mind pricing what else its resources could be doing. Shenhav, Botvinick and Cohen (2013) proposed that the brain allocates control to maximise its expected value: the payoff of exerting control minus its cost.

On this account the weights are largely affective rather than calculated on paper. Damasio (1994) described patients with damage to the ventromedial prefrontal cortex whose reasoning was intact and who could not reach decisions, because the bodily signals that mark one option as heavier than another were missing.

Proposition 1. An action is taken when the perceived weight of staying still exceeds the perceived weight of moving.

The proposition is testable only if weight is measured apart from the action it predicts, through ratings of perceived cost taken before a choice. Prediction 1 in section 10 states that test.

The word perceived carries the argument. A five-minute request for help can weigh more than months of slow loss, because the request carries an immediate social and identity price while the loss arrives later and diffusely. Steel (2007), in a meta-analysis of procrastination, found task aversiveness, delay of reward and impulsiveness among its strongest correlates, which is consistent with the same structure: near costs weigh more than distant ones.

On this account, high agency is the observed pattern of a person for whom staying still is reliably heavier than moving, across many situations. It is an outcome of the weights, and the weights are where intervention is possible.

3. What makes staying still heavier: push

The first source of weight is a gap. Alfred Adler (1927) placed feelings of inferiority at the centre of human striving: the sense of falling short is the spur. Higgins (1987) formalised this as self-discrepancy, the distance between the actual self and the ideal or ought self, and showed that different discrepancies produce different emotions and different motivation. We call this gap the anchor, and the colloquial name for it is something to prove.

A gap needs a reference. Festinger (1954) argued that humans evaluate their abilities by comparison with similar others. Leary, Tambor, Terdal and Downs (1995) proposed that self-esteem works as a sociometer, a gauge of one's standing with others who matter. Girard (1961) went further: desire is copied from a model close enough to compare with, and the model's movement gives the desire its direction.

Proposition 2. Push requires both a gap and a moving reference. If the reference stops moving or is lost, push loses direction even when capacity is unchanged.

Proposition 3. Push can be reduced without any achievement. Any reinterpretation that shrinks the perceived gap, including frameworks that recast status as a social construction, lowers the weight of staying still and so lowers action. The same reinterpretation can also lower the identity cost of failure and so lighten moving; which effect dominates is an empirical question.

4. What makes moving lighter: pull

The second source works on the other side. Panksepp (1998) identified a SEEKING system in mammals, an urge to explore that fires without needing a deficit. Deci (1971) first showed that external reward can undermine intrinsic motivation, and Ryan and Deci (2000) set out self-determination theory, in which sustained intrinsic motivation depends on autonomy, competence and relatedness. Pull makes moving light.

Pull is fragile in a specific place. Hidi and Renninger (2006) described interest in four phases: triggered, maintained, emerging and well developed. Novelty triggers interest, and much interest is lost before it is maintained, plausibly because novelty fades before competence grows enough to be rewarding in itself. External pressure can also crowd pull out. Lepper, Greene and Nisbett (1973) found that preschool children who drew for an expected reward later chose to draw less in free time than children who were never promised one.

5. Identity sets the weight of failure

Much of the weight on the moving side is what a failure would imply about the self. Swann, Stein-Seroussi and Giesler (1992) found that individuals seek confirmation of their self-views even when those views are negative. Oyserman (2009) showed that the same difficulty is read as a sign of importance when a task feels congruent with one's identity, and as a sign of impossibility when it does not. Mueller and Dweck (1998) found that fifth-grade children praised for intelligence persisted less after failure than children praised for effort: the praise had tied their identity to the outcome.

Proposition 4. The single largest weight on the moving side is identity cost. A person whose identity is not at stake in an action finds it light, which is observed as confidence. A person whose identity is fully at stake finds even a small action heavy, which is observed as procrastination, insecurity or low agency.

On this account, behaviour change and habit formation can fail where an identity shift succeeds. A habit lowers effort cost. An identity shift lowers the cost of failure, which we expect is usually the heavier of the two.

6. Two conditions: energy and environment

Energy conditions every weight. Rimfeld, Kovas, Dale and Plomin (2016) estimated the heritability of the perseverance facet of grit at about 37 per cent, and Treadway and colleagues (2012) linked dopamine function in the striatum and ventromedial prefrontal cortex to the willingness to expend effort for reward. Sleep, nutrition and fitness add to or subtract from the same capacity. We discussed the energy side in Adaptability Is Half the Answer.

Environment sets the prices. Social sanction raises the cost of some actions and removes others altogether; the absence of a safety net raises the cost of staying still. Read this way, the components in popular accounts, together with one this model adds (the last row), are weights rather than traits:

7. Decoding the vocabulary

Once the parts are named, the common vocabulary of drive can be restated in terms of them. Each row is the model's restatement, not an established finding.

The list is short because the parts are few. That is the central claim of this paper: the many words for drive describe a small number of mechanisms.

8. Implications for children

Whitehead (1929) described learning as a rhythm of romance, precision and generalisation: wonder first, then disciplined practice, then a deeper interest that returns once the learner is competent. The middle phase is where pull is most likely to fail, because novelty is gone and competence has not yet arrived.

Four implications follow from the model, each a hypothesis to test. Structure such as a fixed time, a teacher and a commitment to someone real carries the heavy middle by keeping staying still heavier. Practice tied to what the child is already pursuing makes moving lighter. Praise directed at process rather than ability keeps identity out of the outcome, as Mueller and Dweck (1998) found. And the reference a child measures against can be chosen deliberately instead of inherited from whoever happens to be nearby.

9. AI and the loss of a reference

Spence (1973) showed that a signal separates individuals only when it is costly to produce in a way that tracks ability. Marks, essays and finished work have served children as references in this sense: they told a child where she stood. Generative AI lowers the cost of producing finished work, so these signals separate children less. If Propositions 2 and 3 hold, the result is weaker push: a child who cannot read where she stands cannot feel a gap. This is a hypothesis, and it is one reason Ilanooo's research records how children direct their own work, which we expect is harder to produce without the ability it reflects.

10. Testable predictions

1. Measures of the perceived cost of acting predict initiation of action beyond trait measures of agency or grit.

2. Removing a comparison reference reduces effort in push-dominant individuals and leaves pull-dominant individuals largely unaffected.

3. Reframing that shrinks a self-discrepancy without any achievement reduces subsequent effort.

4. Interventions that lower identity cost, such as process praise, raise persistence after failure more than interventions that raise effort tolerance.

5. In classrooms where AI-assisted work is hard to tell apart, students report less clarity about their standing and lower effort on tasks that are not assessed.

11. Limitations

This is a synthesis, not new data. The constructs come from traditions that measure differently, and perceived weight is hard to observe directly. The division between push and pull is a simplification: Ryan and Deci (2000) describe a continuum along which external pressure becomes internalised. Whether push or pull dominates in most adults is an open question that this model raises and does not answer. Behaviour scientists already plan change with COM-B, which separates capability, opportunity and motivation (Michie, van Stralen and West, 2011). The contribution here is narrower: placing identity and standing inside a cost model of action, and using it to decode the trait words now applied to children.

Conclusion

High agency is a useful description and a poor explanation. Underneath it, and underneath grit, discipline and motivation, the evidence reviewed here is consistent with one comparison: action follows when staying still is perceived as heavier than moving. Push adds weight to staying still, pull lightens moving, identity prices failure, and energy and environment set the conditions. Each of these parts can be changed, which the trait word cannot.

References

• Adler, A. (1927). Understanding Human Nature.

• Bandura, A. (1977). Psychological Review, 84(2).

• Credé, M., Tynan, M. C., and Harms, P. D. (2017). Journal of Personality and Social Psychology, 113(3), 492-511.

• Damasio, A. (1994). Descartes' Error: Emotion, Reason, and the Human Brain.

• Deci, E. L. (1971). Journal of Personality and Social Psychology, 18(1).

• Duckworth, A. L., Peterson, C., Matthews, M. D., and Kelly, D. R. (2007). Journal of Personality and Social Psychology, 92(6).

• Eccles, J. S., et al. (1983). Expectancies, values, and academic behaviors. In J. T. Spence (Ed.), Achievement and Achievement Motives.

• Festinger, L. (1954). Human Relations, 7(2).

• Girard, R. (1961). Deceit, Desire and the Novel (English edition 1965).

• Hidi, S., and Renninger, K. A. (2006). Educational Psychologist, 41(2).

• Higgins, E. T. (1987). Psychological Review, 94(3).

• Kelley, T. L. (1927). Interpretation of Educational Measurements.

• Kurzban, R., Duckworth, A., Kable, J. W., and Myers, J. (2013). Behavioral and Brain Sciences, 36(6).

• Leary, M. R., Tambor, E. S., Terdal, S. K., and Downs, D. L. (1995). Journal of Personality and Social Psychology, 68(3).

• Lepper, M. R., Greene, D., and Nisbett, R. E. (1973). Journal of Personality and Social Psychology, 28(1).

• Lewin, K. (1936). Principles of Topological Psychology.

• Lewin, K. (1951). Field Theory in Social Science.

• Michie, S., van Stralen, M. M., and West, R. (2011). Implementation Science, 6, 42.

• Mueller, C. M., and Dweck, C. S. (1998). Journal of Personality and Social Psychology, 75(1).

• Oyserman, D. (2009). Journal of Consumer Psychology, 19(3).

• Panksepp, J. (1998). Affective Neuroscience: The Foundations of Human and Animal Emotions.

• Rimfeld, K., Kovas, Y., Dale, P. S., and Plomin, R. (2016). Journal of Personality and Social Psychology, 111(5).

• Rotter, J. B. (1966). Psychological Monographs, 80(1).

• Ryan, R. M., and Deci, E. L. (2000). American Psychologist, 55(1).

• Shenhav, A., Botvinick, M. M., and Cohen, J. D. (2013). Neuron, 79(2).

• Spence, M. (1973). Quarterly Journal of Economics, 87(3).

• Steel, P. (2007). Psychological Bulletin, 133(1).

• Swann, W. B., Stein-Seroussi, A., and Giesler, R. B. (1992). Journal of Personality and Social Psychology, 62(3).

• Treadway, M. T., et al. (2012). Journal of Neuroscience, 32(18).

• Whitehead, A. N. (1929). The Aims of Education and Other Essays.