Felt security over time is the slow, cumulative softening of your worry about where you stand, as month after month of a responsive partner quietly rewrites the insecure assumptions you arrived with into something closer to earned securityearned securityReaching a secure, trusting way of relating despite a painful past, shown by narrating hard experiences reflectively and coherently rather than by having had…Glossary →.
Attachment theory holds that we carry internal "working modelsinternal working modelA mental template of yourself, other people, and relationships, formed in early life and carried into adult bonds, that quietly shapes how you read…Glossary →" — implicit expectations about whether others will be available and whether we're worth responding to 1 — that bias how we read ambiguous moments; the same unanswered text reads as "they've lost interest" or "they're busy" depending on the model. Felt securityfelt securityThe settled background sense that a partner is reliably there for you, so your nervous system no longer has to keep asking whether they are.Glossary → is the subjective end of this, the appraised sense of safety Sroufe and Waters named 2. This flag is about that sense changing across time — insecure models revised toward security through repeated evidence, what researchers call earned security 3. It is not the partner's behavior (#26, #103) or a present-tense safe harbor (#93, #100); it is the change in you, measured in the difference between who you were at month three and at year three. The feeling finally catching up to the facts.
Arriaga and colleagues' Attachment Security Enhancement Model lays out how relationships revise underlying models, not just buffer the moment — anxiety declining where confidence in the self grows, avoidanceattachment avoidanceAn attachment pattern marked by discomfort with closeness — pulling away when intimacy rises and managing that unease through distance rather than through…Glossary → where positive dependence grows 4, with longitudinallongitudinalA study that follows the same people over time (often years), allowing researchers to see change and sequence, not just a single-moment snapshot.Glossary → support from self-efficacy predicting declining anxiety 13. Models can change, but not easily: Fraley's meta-analysis finds real stability and room for revision 511, and a large dyadic study found insecurity only partially changes, mostly in anxiety 12. Earned security is genuine but methodologically contested 67. The proximate engine is perceived partner responsivenessresponsivenessA partner's demonstrated understanding, validation, and care — how much you feel truly seen and supported, which research treats as a core driver of intimacy.Glossary → 8 — with the catch that insecure models under-read the very responsiveness that would revise them, which is why change is slow 8. Security also broadens the capacity for exploration and care 9, the premise EFT is built on 10.
Why it matters
Why it matters. The dividend is a lower resting vigilance: energy once spent scanning for abandonment becomes available for living, and a hard argument reads as a problem inside a stable bond rather than a threat to it.
How to cultivate it
How to cultivate it. Supply consistent, legible responsiveness, name the disconfirming evidence on purpose, allow dependence, and measure progress in years, not weeks.
The honest caveat
The honest caveat. The evidence is real but shouldn't be oversold. The enhancement model is an influential synthesis, not proof any specific model will change 4; adult attachment is substantially stable with revision layered on top 512; the earned-secure construct is contested 67; and causality is entangled, since more secure people may elicit more responsiveness 8. Treat this as a well-grounded direction of travel, not a guarantee.