Metacognition of Ignorance: What Can Animal Studies Teach Us?

Metacognition pertains to the ability to reflect on one’s mental states, and specifically to knowing that one knows. When children are directly asked whether they know, e.g., what is inside a box, they find it surprisingly difficult. They often answer with a guess but are then unable to say whether they are merely guessing or really know before the age of 4 or 6 years. Do they have merely problems with the semantics of mental vocabulary or are they genuinely oblivious about what they know or don’t know?

                Semantic problems are supported by the fact that much younger children and many species of animals do react differently when they know than when they are ignorant. For instance, when ignorant or uncertain about the correct response they search for more information before deciding or they opt out of making a decision altogether. However, this raises deeper conceptual issues of whether these behaviours unambiguously depend on reflecting on one’s mental states (metacognition). I discuss three problems with this interpretation. One pertains to the learning procedure in animal studies. The other two address deeper conceptual problems of what constitutes metacognition.

                The base-line conditioning problem. Most animal studies on metacognition work with the opt-out paradigm. One explanation, not considered so far, is that animals acquire the tendency to press the opt out key. This response tendency is overruled only if the discriminate stimulus value triggers a stronger alternative response. This basic conditioning explanation can account for the difference in response and for generalisation of the opt-out response to novel discriminatory stimuli.          

                The state/content problem. All relevant evidence is based on finding that a different response is given when one is certain of (knows) what to do than when one is uncertain (ignorant). The open question is whether this difference in behaviour necessarily requires reflecting on one’s ignorance (having a mental state with the content “I am ignorant”) or whether simply being in the state of ignorance is sufficient.

                The internal/external problem. Another general problem of interpretation occurs due to a bijective mapping of internal states to external situations. If I freeze when seeing a snake, is it because I am afraid (a mental state) or because I am facing a frightening stimulus? Similarly, do I opting out when being uncertain because I am aware of being uncertain or is it because I am faced with an unclear (difficult) stimulus.