TABLE 3.

Summary of infant studies discussed in the section “Generalization”.

Paper	Age	Paradigm	Main conclusion
Change in stimulus
Milewski and Siqueland, 1975	1 month	Non-nutritive sucking	From very early on, infants can generalize their learning of a contingency to a new stimulus when the change occurs immediately after learning.
Hartshorn et al., 1998 review Rovee-Collier, 1999 review	2 to 12 months	Mobile paradigm Train paradigm	Before 6 months, infants cannot generalize their learning of a contingency to a new stimulus when the change occurs 1 day after learning. After 6 months, infants can generalize their learning of a contingency to a new stimulus when the change occurs within 2 weeks, but not for longer delay.
Butler and Rovee-Collier, 1989	3 months	Mobile paradigm	Young infants cannot generalize their learning of a contingency to a new stimulus and context when the change occurs 1 day after learning.
Fagen and Rovee, 1976 Mast et al., 1980	3 months	Mobile paradigm	Young infants can generalize their learning of a contingency when quantitative aspects of feedback are modified immediately or 24 h after learning.
Fagen et al., 1976	3 months	Mobile paradigm	A change of more than two objects in the mobile prevents generalization.
Hayne et al., 1986	2 months	Mobile paradigm	A change of more than one object in the mobile prevents generalization.
Adler and Rovee-Collier, 1994	3 months	Mobile paradigm	A change in appearance of the elements of the mobile (painted letters) prevents generalization.
Change in context
Hartshorn et al., 1998 review Rovee-Collier, 1999 review	2 to 12 months	Mobile paradigm Train paradigm	Infants are able to generalize their learning to a new crib when tested 1 day after training before 6 and up to one month after 6.
Hayne et al., 1991	3 months	Mobile paradigm	A change of room prevents generalization when testing occurs 2 weeks after training but not when it occurs 1 day after training.
Hartshorn and Rovee-Collier, 1997	6 months	Train paradigm	A change of room prevents generalization when testing occurs 3 weeks after training and is preceded by a reactivation the day before.
Rovee-Collier and Dufault, 1991	3 months	Mobile paradigm	Training in multiple contexts overrides the lack of generalization to a new context.