TALE OF THREE FUNCTIONS
cial functions of AM. For example, theories of socio–emotional pro- cesses across the lifespan suggest that initiating relationships is particularly important in young adulthood, but that selectively maintaining close social bonds is more the norm in later life (Carstensen, 1993). Our reliance on a younger adult sample in this study may have overrepresented the distinct importance of forging new relationships.
These two social factors showed differing relations with the rele- vant RFS subscales. Developing Relationships did not show the same clear relation to expected social indicators as did Nurturing Re- lationships. Nurturing Relationships clearly was related to the RFS–Conversation subscale and also to an overall tendency to talk about one’s past with other people. Nurturing Relationships also was the most highly endorsed subscale. On average, people reported using AM to nurture relationships “occasionally” to “often.” Thus it appears that the most central social function of AM is social bonding in existing relationships. How and how often individuals use AM to initiate or develop new relationships needs to be followed up in fu- ture work to understand if it is indeed a separate, and well–utilized, function of AM.
LIMITATIONS OF THE CURRENT STUDY
The current study examines three broad functions of AM that have, to this point, received largely theoretical attention. Our method was to have individuals simply report on their uses of AM to serve a vari- ety of ends. The accuracy and validity of such reports requires that people must be aware of, and insightful about, how they use AM. Fu- ture research needs to examine more fully the extent to which people are able to reflect in this manner. Moreover, there may be other func- tions of AM of which individuals are not aware, and there may be in- dividual differences in people’s awareness of the extent to which memory serves important functions in their everyday life. Thus, we believe that our data is reasonable for what it is, but it clearly can only assess the functions of AM that individuals know or can recognize that they use. Other methods that assess function less directly could add more pieces to this developing picture.
One possible critique of these findings is that the overall factor so- lution only accounted for 50% of the variance. However, this was our first attempt to assess the theoretical AM function constructs. Given