Two experiments involving a training paradigm examined the role of adu
lt head movement in early gaze following. In the first experiment, thr
ee groups of 9-month-old infants interacted with an adult, who, over a
series of trials, presented one of two cues for the appearance of an
interesting sight to the side. Two of the groups (nonspontaneous) were
composed of infants who were not engaging in gaze following at the st
art of the session. One group saw the adult produce a head-rum movemen
t to the side; the other saw only the static head orientation at the e
nd of the same movement. The third group (spontaneous), composed of in
fants who were engaging in gaze following at the start of the session,
also saw the adult's static head orientation. Results showed that non
spontaneous infants could learn to follow the dynamic head turn, but n
ot the static head orientation. Spontaneous infants did follow the sta
tic head orientation. Given this result, a further group of nonspontan
eous 9-month-olds were tested. Using the same paradigm, these infants
observed the head-turn movement without the final static head orientat
ion. They also followed gaze at above chance levels. The second experi
ment involved the same design and procedure as the first, with the exc
eption that a head-tilt movement was substituted for the static head o
rientation. Again, nonspontaneous infants acquired gaze following with
the head rum, but not with the head tilt. Spontaneous infants were ab
le to learn the signal value of the head tilt in the experimental sess
ion. A further group of nonspontaneous 9-month-olds was tested in the
same paradigm. These infants observed an adult head turn that stopped
before reaching the target at 30 degrees. These infants also were able
to learn to turn toward the correct target. These results are discuss
ed in the context of the developmental origins of the gaze-following r
esponse necessary for joint visual attention and the infant's understa
nding of attention.