Source
- Deci & Ryan (2000), LIT-558, read in NOTE-439: “Internalization and Need Satisfaction” (pp. 237–238), and need thwarting and its accommodations (pp. 247–251).
- Ryan & Deci (2000), LIT-559, read in NOTE-442: “Facilitating Integration of Extrinsic Motivation” (pp. 73–74).
What was actually shown
The claim as SDT states it. “Supports for relatedness and competence thus facilitate internalization and can be sufficient to produce introjected values or compartmentalized (poorly integrated) identifications. However, for a regulation to become more integral to one’s self, supports for autonomy are also required” (LIT-558, p. 237). “Contexts can yield external regulation if there are salient rewards or threats …; contexts can yield introjected regulation if a relevant reference group endorses the activity …; but contexts can yield autonomous regulation only if they are autonomy supportive” (LIT-559, pp. 73–74).
The experiment. Deci, Eghrari, Patrick and Leone (1994), as both reviews report it: people did a dull task with zero to three facilitating factors (a meaningful rationale, acknowledgement that the task is dull, choice rather than control). Everyone internalized something. With two or three factors, later free-choice behaviour correlated positively with reported valuing, enjoyment and freedom; with zero or one, negatively: “people who behaved more felt less free and enjoyed the activity less” (LIT-558, pp. 237–238). The authors read the second pattern as introjection.
Field studies, correlational. Children whose parents were autonomy supportive internalized school values more (Grolnick & Ryan 1989); autonomy supportive religious parents produced identified rather than introjected belief (Strahan 1995); medical students of autonomy supportive instructors internalized course values, visible six months later (Williams & Deci 1996) (LIT-559, p. 74; LIT-558, pp. 237, 240). Contingent parental regard produced introjected regulation (Assor, Roth & Deci, LIT-558, p. 248).
What this does not say
- It does not say history alone decides. SDT also holds that internalization can be revisited “at any point along this continuum” (LIT-559, p. 73): a regulation introjected once can be integrated later. The claim is about how acquisition conditions shape the kind of internalization, not that it is fixed for life. The title’s “shapes” is about the state produced, not its permanence.
- It does not handle manipulation. A manipulator who supplies rationale, acknowledgement and apparent choice would, on this account, produce integration. SDT’s papers do not discuss that case.
- The experiment is held second-hand. Deci et al. (1994) is not filed.
Connections
- Extends THEORY-054. That account says autonomy is degree of integration; this one says what produces integration rather than introjection. Same mechanism, widened to its causes.
- THEORY-040 (snapshot conditions are met by a manipulated agent, and no historical condition in the record escapes). This is a historical condition with experimental support, the kind THEORY-040 finds only in philosophy (Christman) and in Pettigrew’s consent account (LIT-202). It sorts histories by autonomy support, which is defined using the notion of autonomy it serves, so it meets the circularity worry THEORY-040 records from LIT-296. It does not show that a manipulated history fails the condition.
Note of 2026-10-29: what holds it back
Kept Proposed by the tenth status review. The core experiment (Deci, Eghrari, Patrick and Leone 1994) is held only as two reviews report it and is not filed, and the decisive test is not in the record: two groups ending with the same present endorsement, acquired under autonomy support or under control, compared on maintenance and well-being.
Note of 2026-11-03: where the history would have to show
CASE-191 is a construction in which a capacity installed from outside and then self-maintained is, in itself, identical to one the agent formed: no knockout after formation tells them apart. If this theory is right, the history of acquisition shapes autonomy anyway, so the difference has to lie in something the construction leaves out, such as how the capacity is integrated with the agent’s other values and demands (THEORY-054). That sharpens what a test of the theory should measure. Not the regulation itself, but its relations to the rest. Deci et al. 1994 (LIT-1973), the primary study this theory holds second-hand, is registered.
The status is unchanged.
Note of 2026-11-03 (later): the readings
Deci et al. 1994 is Unavailable, so the experiment is still held second-hand. Ryan et al. 2022 (NOTE-1500) report Steingut et al. 2017, 23 rationale experiments with perceived autonomy at r = .20: the nearest meta-analytic evidence, but it measures present endorsement, not the history against the structure that CASE-191’s note asks a test to separate. Chirkov et al. 2003 (NOTE-1513) hold internalization predicting well-being across four cultures, correlationally.
The status is unchanged.