Cognitive Daily

Over at Developing Intelligence, Chris Chatham has a fascinating discussion of infantile amnesia, which he tantalizingly terms a “myth.” Chris cites research demonstrating that infants can and do remember things, even stories read to them in the womb:

3-day-old infants were capable of distinguishing a particular passage (from Dr. Seuss’s “Cat in the Hat”) that had been read to them twice daily for the last 6 weeks of gestation from similar passages (matched for word count, length, and prosody). What’s more, these infants preferred the familiar passage even if spoken by someone other than their mother, strongly suggesting that they had encoded (and retained) a relatively high-level representation of the passage’s auditory content.

Wow! But of course, this sounds more like a conditioned response than a “memory” in the everyday sense (though arguably there’s no difference: if we’ve learned something, then it’s a memory, right?). After citing additional research on long-term memory, Chatham reaches a fascinating conclusion:

In summary, there is some reason to believe that the phenomenon of infantile amnesia has been widely mischaracterized. Infants show surprisingly robust long-term memories, and both the explicit and implicit memory systems appear functional very early in life. Based on this evidence, it seems unlikely that retention failure underlies the apparent loss of all early-life experiences. Rather than being completely forgotten, our earliest experiences may actually be mislabeled.

This conclusion is supported by some work we’ve discussed here on CogDaily where toddlers were shown a “magic” machine and then asked about it six months to a year later:

None of the children interviewed used any of the words that they did not know at the time of the original demonstration to describe their memory of the event. Though they clearly could remember the experience, and even showed the experimenters how the machine worked, they didn’t use the proper words for the parts of the machine (“handle,” “knob”) if they hadn’t known them at the time of the original event. The memory existed, but the words were not associated with the memory.

Simcock and Hayne argue that these memories simply are not ever encoded in language, and for that reason, never become part of an adult’s autobiographical memory.

In other words, the memories are there; they just aren’t encoded in a way that makes them easy to retrieve later.


  1. #1 speedwell
    January 18, 2007

    When I had my first child, all sorts of memories of myself as a very young child came back to me in the first 18 months of his life. They were there, but needed something, possibly some sort of empathy with my son, to trigger their recall. Other first-time mothers have told me they have experienced the same thing, some as far back as late in pregnancy.

  2. #2 Dave Munger
    January 18, 2007

    Wow, that’s amazing! If I hadn’t just read the research myself, I would have thought you and your friends had lost your minds….

  3. #3 Gordon Worley
    January 18, 2007

    Interesting results. This is a subject that has always bugged me, because I’ve never heard a satisfactory explanation for it, just rationalizations about “unneeded” or “harmful to development” information. Even suppositions that the memory system isn’t in place are somewhat suspect, since anyone who’s spent much time with an infant knows they have long term memories, even for sporadic events, like occasional visitors. The problem of encoding and retrieval sounds fits the evidence much better, in my opinion.

  4. #4 Chris Chatham
    January 18, 2007

    I’m surprised that the Hayne paper I read didn’t cite the Simcock & Hayne paper you mention here. Wonderful addition!

    That research really underscores the importance of language as a retrieval cue, and suggests that perhaps synaptic pruning is not such a big player as I suggested (in comments).

    Anyway, thanks for pointing this out.

  5. #5 Dana Leighton
    January 18, 2007

    But of course, this sounds more like a conditioned response than a “memory” in the everyday sense (though arguably there’s no difference: if we’ve learned something, then it’s a memory, right?).

    ACK! that’s a tangled web. Well, with the exception of classical conditioning, I think you’re right. Calling a classically conditioned response a “memory” (as we typically use the term) is a bit of a stretch I think. It is a biological reflexive response. For operant conditioning, I think using the term “memory” is perhaps more appropriate, since we have to “remember” (or better, use the terminology of operant conditioning: “associate”) that a particluar behavior previously resulted in a reinforcing or punishing consequence.

    I looked up the Hayne article Chris Chatam cited. The 3-day-old infants were operantly conditioned to change their sucking rate, using the story as a reinforcement. In this case, responding to the story was not operantly conditioned prior to birth, but rather, the story read to them in gestation apparently became a reinforcing stimulus.

    So, yes, the sucking rate behavior was operantly conditioned, but the familiar story was apparently “remembered” as a desirable stimulus. The contingencies of the environment (administration or withholding the famiiiar passage) were arranged to occasion the altered sucking rate. But, the familiar story was somehow a pleasurable stimulus, which seems to imply it was encoded in memory which persists at least 2-3 days after birth.

    I think the more interesting hypothesis for childhood amnesia is that children prior to 3-4 years old have different schemas for their worlds, and that there is such a profound shift in schema formation and maintenance at about 3-4 years old, that adults cannot access information encoded using the previous schemata. There is some indication that the more parents talk with children, and help them learn how adults structure their schematic representations of their worldview, and might allow earlier recall of episodic or autobiographical memories.

    For example, see:
    Mullen, M. (1994). Earliest recollections of childhood: A demographic analysis. Cognition, 52, 55-79.

  6. #6 David Group
    January 19, 2007

    Jay Leno has occasionally mentioned that he remembers being breast-fed as an infant. I had always been inclined to dismiss this as a false memory (similar to Paiget’s “memory” of witnessing his nanny being mugged.) Now, I’m not so sure.

  7. #7 Ken C.
    January 20, 2007

    Chris Chatham writes:

    at least in some ways, infantile amnesia may actually be a myth

    which is a puzzling claim as stated, at least as far I understand the term “infantile amnesia”. Surely he is not claiming that most people think they don’t have any recollection of being one year old, but in fact they do? And I don’t think anyone thought that all learning and skills acquired in infancy are simply lost: I learned to walk and talk in infancy, for example, and I still manage to do both, to some degree. I’m not sure that I remember a whole lot more specific events from when I was, say, ten years old, than from when I was four. Why is it surprising that the fleeting images and moments that constitute my memory of any age are weak or absent for ages zero to two?

  8. #8 David Group
    February 2, 2007

    Have any studies on infantile amnesia been done with individuals who have language disabilities (e.g., dyslexia)?

New comments have been disabled.