What Are Autobiographical Memories? A Reply to Bauer,Baker-Ward,Krøjgaard,Peterson,and Wang (2019)

Consider the following account of a childhood memory: “I really wanted to say something to my mother but had not yet learned to talk.” The person who recalled this memory was more than 30 years old but dated the memory to below the age of 1 year. But is it really possible to remember not being able to talk? Consider several more accounts of memories (Akhtar, Justice, Morrison, & Conway, 2018b):

Sitting in a pram with my younger brother he was a very small baby at the time, I am 14 months older than my brother so I must have been 2 or younger. I remember being in the pram the rain cover was on and my brother was under a white blanket. I remember having a feeling of warmth and security. (age 2 months; age at recall: 46–50 years)

I have a good Photographic memory which may explain why I remember a certain part in the past when I was only 2 months old. I am now 33 years old. Memory: I can see myself positioned outside my parents dark green Morris Minor (with dark wood trims looking about 4 meters facing the back of the car). The car is on the left hand side of a road, weather is bright, colours are vivid. Tall grass either side of the road. I think the back doors are shut. I can see no one. When mentioned to my parents, they thought it was impossible at the age I was at the time, told as being 2 months old.

Sitting on my “potty” listening to the radio (Listen With Mother) while my own mother took my brother to school. (age 2 years or below; age at recall: 41–45 years)

Could a 2-month-old know what a Morris Minor is or what a rain cover is? Is it possible to remember the radio program one was listening to at age 2? The evidence suggests that it is not, which is why we argued that these accounts are best described as fictional memories (Akhtar, Justice, Morrison, & Conway, 2018a, pp. 1616–1617). From our perspective, such reports are part of the life story formed by the self-memory system (Conway, Justice, & D’Argembeau, in press; Conway & Pleydell-Pearce, 2000), a major part of which is autobiographical memory. Autobiographical memory consists of conceptual knowledge of one’s life represented at different levels of specificity and of episodic memories that contain fragmentary details from the time near the experience being recalled (Conway, 2009). Indeed, this is what we meant by the term “full autobiographical memories” (Akhtar et al., 2018a, p. 1612). Episodic memories are then event-specific and are embedded in experience-distant associated networks of conceptual knowledge. From our perspective, infants and young children do not have autobiographical memories because they have yet to develop the conceptual-knowledge networks in which fragmentary episodic memories will eventually become embedded. Younger children, although lacking full autobiographical memories, may well have mental representations that at least approximate episodic memories.

One critically important aspect of conceptual knowledge is the development of the self. Indeed, Bauer et al. acknowledged that “there is consensus that the memory must be of a specific past event and that it be self-referential” (p. 1398, emphasis added). As most people would also agree, this self does not emerge until around 2 years of age (of course, there are large individual differences here and, in any case, the development of the self is likely a process extended in time). Nevertheless, and to clarify, our point is that to the extent that the self does not exist or exists only in rudimentary form there can be no full autobiographical memories (e.g., Howe, 2000; Howe & Courage, 1993, 1997; Howe, Courage, & Edison, 2003; Ross, Hutchison, & Cunningham, 2019).

Another issue here is what is meant by “memory . . . of a specific past event.” We know from numerous studies (see Conway, 2005, 2009) that memories of specific events (i.e., episodic memories) are fragmentary, time compressed, and contain many unconscious and conscious inferences. But the schemas and conceptual networks that support such filling in of episodic memory details are only just taking shape in infants and young children and are hardly present, if at all, in neonates. Thus, observations that an infant can recognize a row of objects, look at a new item in a display, even recall some detail, are hardly surprising. This is the beginning of the development of episodic memory but not yet full autobiographical memory itself.

Our view of autobiographical memory differs from that of Bauer et al. in that we believe that what fragments remain in memory from early experiences become integrated into the life story, making them consistent with current beliefs. During the process of integration into the life story, memory fragments may be altered to fit. Thus, the fact that infants and young children show evidence of memory does not mean that they have full conceptually rich autobiographical memory. It follows from this that adults who report conceptually rich autobiographical memories (such as those listed above) dating to below 3 years of age are giving accounts that we believe are highly embellished (see Howe, 2013). Such embellishments may be based on episodic details retained from these very early years, conceptual knowledge of their infancy acquired later, family stories, and so forth. Whatever the case, they are fictional in the way described by Akhtar et al (2018a). In agreement with other research (e.g., Bruce, Dolan, & Phillips-Grant, 2000; Bruce et al., 2005), our view is that representations approaching the complexity and richness of adult autobiographical memory do not generally begin to emerge until the age of 5 to 6 years (see Wells, Morrison, & Conway, 2014). Of course, and again, there are large individual differences in this, although we believe that the emergence of a range of full, conceptually rich, autobiographical memories dates to later rather than earlier childhood and infancy.