Metacognition: Knowing about Knowing

by Janet Metcalfe (Editor), Arthur P. Shimamura (Editor)

Metacognition offers an up-to-date compendium of major scientific issues involved in metacognition. The twelve original contributions provide a concise statement of theoretical and empirical research on self-reflective processes or knowing about what we know. Self-reflective processes are often thought to be central to what we mean by consciousness and the personal self. Without such processes, one would presumably respond to stimuli in an automatized and environmentally bound manner--that is, without the characteristic patterns of behavior and introspection that are manifested as plans, strategies, reflections, self-control, self-monitoring, and intelligence. A Bradford Book

Genres: Philosophy, Psychology

Paperback

First published April 7, 1994

Reviews

[G · Rating 5 out of 5]

A principios de los años 2000 este libro era el mejor estado del arte sobre metacognición, que no es otra cosa que la introspección, pero en versión actualizada con criterios de las ciencias cognitivas. Lo interesante es la recursividad. Es pensamiento sobre el propio pensamiento. Lo mejor es la fuerza epistemológica de los experimentos que ya se habían hecho por aquel entonces. Pude hacer buena parte del marco teórico de mi tesis doctoral sobre esta idea de recursividad y sus límites computacionales en el sesgo mental de ilusión de verdad. Es un libro lúcido y bien escrito. Todavía persigo preguntas que nacieron de esta lectura.

[Max · Rating 3 out of 5]

Read for an overview of what the psychological research on metacognition is up to, and noticed saddingly late that the book is almost as old as I am.

I enjoyed getting a glimpse of the historical development of the research and a rough idea what people can and cannot reliably predict about their own performance. When, for example, physics and music students in one study read about both concepts from music and physics, they could broadly predict that they will perform better in questions about concepts from their respective field, unsurprisingly. But their predictive accuracy* how well they would perform on individual questions was very low, to my surprise. I would've thought that you can at least predict a little bit how well you can apply recently aquired knowledge, but apparently this was not the case here. [2]

I liked the proposal that a major part of problem solving lies in useful problem representations.
"One advantage is that good representations allow the problem solver to organize blocks of planned moves or strategies as a single "chunk" of memory. In other words, good representations help reduce the memory demands found in many problems. Second, good representations allow the problem solver to organize the conditions and rules of a problem and to determine whether certain steps are allowable and productive. Finally, good representations allow the problem solver to keep track of where he or she is in terms of reaching a solution and to foresee potential obstacles to reaching the solution."
I would like to know more about the state of the art of how one might go about modeling representation learning.

Also, I think everyone today will note how early days the theory side of the research still was. The most concrete models where abstract flowcharts with "metacognition" and "object level cognition" represented by black boxes. I'm looking forward to look at Dunlosky and Metcalfe's 2008 textbook on metacognition and see what happened in the 14 years of research.


* The predictive accuracy here was measured with Goodman & Kruskal's Gamma. This measure has the downside of mixing together the ability to distinguish between correct and incorrect performance, and a bias towards over- or underconfidence. More on this in the excellent review, "How to measure metacognition" by Fleming and Lau [1].

[1] https://www.frontiersin.org/articles/...
[2] https://link.springer.com/article/10....