A learning theory of attachment: Unraveling the black box of attachment development

Guy Bosmans*, Marian J. Bakermans-Kranenburg, Bram Vervliet, Martine W.F.T. Verhees, Marinus H. van IJzendoorn

*Corresponding author for this work

Research output: Contribution to JournalReview articleAcademicpeer-review


Attachment is an inborn behavioral system that is biologically driven and essential for survival. During child development, individual differences in (in)secure attachment emerge. The development of different attachment behaviors has been traditionally explained as a process during which experiences with (lack of) responsive and supportive care are internalized into working models of attachment. However, this idea has been criticized for being vague and even untestable. With the aim of unraveling this black box, we propose to integrate evidence from conditioning research with attachment theory to formulate a Learning Theory of Attachment. In this review, we explain how the development of individual differences in attachment security at least partly follows the principles of classical and operant conditioning. We combine observed associations between attachment and neurocognitive and endocrinological (cortisol, oxytocin, and dopamine) processes with insights in conditioning dynamics to explain the development of attachment. This may contribute to the explanation of empirical observations in attachment research that are insufficiently accounted for by traditional attachment theory.

Original languageEnglish
Pages (from-to)287-298
Number of pages12
JournalNeuroscience and Biobehavioral Reviews
Early online date8 Apr 2020
Publication statusPublished - Jun 2020


  • Adaptation
  • Attachment
  • Conditioning
  • Contingency
  • Cortisol
  • Development
  • Dopamine
  • Information processing
  • Oxytocin
  • Secure base script
  • State
  • Trait


Dive into the research topics of 'A learning theory of attachment: Unraveling the black box of attachment development'. Together they form a unique fingerprint.

Cite this