ArticlePDF Available

Developing a Reflective Mind: From Core Metacognition to Explicit Self-Reflection



Metacognition is the ability to monitor and control cognition. Because young children often provide inaccurate metacognitive judgments when prompted to do so verbally, it has long been assumed that this ability does not develop until late childhood. This claim is now challenged by new studies using nonverbal paradigms and revealing that basic forms of metacognition—such as the ability to estimate decision confidence or to monitor errors—are present even in preverbal infants. This new line of evidence suggests that young children adapt to their environment not only by considering their physical and social surroundings but also by reflecting on their own cognitive states.
Current Directions in Psychological
2019, Vol. 28(4) 403 –408
© The Author(s) 2019
Article reuse guidelines:
DOI: 10.1177/0963721419848672
Metacognition is our ability to reflect on our own mental
representations to regulate cognition and optimize learn-
ing. Given metacognition’s importance for education, its
development has been studied intensively since the
1970s. For several decades, on the basis of studies that
primarily relied on verbal reports, researchers argued
that young children have poor and inefficient self-
reflective abilities (for reviews, see Flavell, 2000; Sodian,
Thoermer, Kristen, & Perst, 2012). Yet it now appears
that young children’s limitations largely reflect an inabil-
ity to provide accurate metacognitive reports rather than
an absence of metacognition per se. Indeed, a new line
of investigation relying on nonverbal paradigms and
focusing on error monitoring (i.e., detecting one’s own
mistake) and decision confidence (i.e., estimating the
probability that a choice was correct) has revealed that
even preverbal infants engage in rudimentary forms of
metacognition. In this article, we briefly review this bur-
geoning literature before proposing a framework that
integrates recent findings with more traditional research
documenting a slow and effortful development of explicit
aspects of metacognition.
Early Forms of Metacognition
Over the last decade, several approaches have been
elaborated to examine the development of metacogni-
tion without relying on verbal reports. A first approach
revealed that when 3- to 5-year-olds are trained to use
a nonverbal confidence scale depicting pictures of a
confident or a doubtful peer, they quickly learn how
to use it appropriately (Ghetti, Hembacher, & Coughlin,
2013). That is, they provide higher confidence judg-
ments for correct compared with incorrect responses,
thereby displaying metacognitive sensitivity (i.e., accu-
rate evaluations of their performances). One limit of
this approach, however, is that nonverbal confidence
scales still require an understanding of the concept of
confidence: To use them appropriately, children must
understand how the symbolic representations depicted
by the scale map onto their internal states of uncer-
tainty. Thus, this type of measure cannot be extended
to test younger children, who cannot be verbally
instructed on how to use such a scale.
Consequently, a second approach consisted of adapt-
ing behavioral methods stemming from the animal lit-
erature. In particular, in the opt-out paradigm (Hampton,
848672CDPXXX10.1177/0963721419848672Goupil, KouiderDeveloping a Reflective Mind
Corresponding Authors:
Louise Goupil, Institut de Recherche et Coordination en Acoustique et
Musique, 1 Place Igor Stravinsky, 75001, Paris, France
Sid Kouider, École Normale Supérieure, Laboratoire de Sciences
Cognitives et Pscyholinguistique, Pavillon Jardin, 29 Rue d’Ulm, 75005,
Paris, France
Developing a Reflective Mind: From Core
Metacognition to Explicit Self-Reflection
Louise Goupil1 and Sid Kouider2
1Science and Technology of Music and Sound, Unités Mixtes de Recherche 9912, Centre National
de la Recherche Scientifique/Institut de Recherche et de Coordination en Acoustique/Musique,
Paris, France, and 2Brain and Consciousness Group, Département d’Études Cognitives,
École Normale Supérieure, Centre National de la Recherche Scientifique, École des Hautes
Études en Sciences Sociales, PSL Research University
Metacognition is the ability to monitor and control cognition. Because young children often provide inaccurate
metacognitive judgments when prompted to do so verbally, it has long been assumed that this ability does not develop
until late childhood. This claim is now challenged by new studies using nonverbal paradigms and revealing that basic
forms of metacognition—such as the ability to estimate decision confidence or to monitor errors—are present even
in preverbal infants. This new line of evidence suggests that young children adapt to their environment not only by
considering their physical and social surroundings but also by reflecting on their own cognitive states.
metacognition, confidence, error monitoring, development, self-reflection
404 Goupil, Kouider
2009), participants are presented with a binary choice,
either in a condition in which they have no option but
to choose by themselves (forced choice) or in a condi-
tion in which they are also given the additional option
to avoid responding (i.e., opt out). The rationale behind
this paradigm is that if participants are able to monitor
their own uncertainty, they should opt out specifically
when they are uncertain, and thereby avoid making
errors in choosing between the two alternatives. In such
a task, 3-year-olds (Balcomb & Gerken, 2008) and even
20-month-olds, who cannot yet verbally communicate
their knowledge states (Goupil, Romand-Monnier, &
Kouider, 2016), have been found to opt out appropri-
ately to avoid errors. In one of these experiments,
20-month-old toddlers had to remember the location
of a toy hidden for variable durations (to induce several
levels of difficulty) before pointing to indicate where
they wanted to recover it (Goupil etal., 2016). Although
toddlers in a control group had no option other than
to decide by themselves, toddlers in a test group were
given the extra option of asking their caregiver for help
through nonverbal communication. In this test group,
only two training trials allowed toddlers to understand
that they could withhold their response when feeling
uncertain and ask their caregiver for help instead.
Toddlers opted out not only to avoid difficult choices
but also to avoid making errors irrespective of task dif-
ficulty; even in easy trials, toddlers in the test group
made fewer errors than toddlers in the control group.
Importantly, the rate of correct choices remained the
same across the two groups, demonstrating that the
toddlers opted out specifically to avoid making errors.
Overall, the results revealed that toddlers were not
merely responding on the basis of risk avoidance (e.g.,
escaping from difficult choices) or other first-order (i.e.,
cognitive) mechanisms. Rather, the results showed that
children appraised their probability of success (i.e.,
asked for help when confidence was low; see Goupil
etal., 2016, for further information).
Converging evidence comes from a third approach
elaborated by comparative psychologists to measure
metacognition while ruling out first-order interpreta-
tions: postdecision wagering. In such a task, 5-year-old
children place higher bets on correct compared with
incorrect responses (Vo, Li, Kornell, Pouget, & Cantlon,
2014). Recently, we used a totally nonverbal variant of
this paradigm to demonstrate that preverbal infants also
display such metacognitive sensitivity (Goupil &
Kouider, 2016). Infants made a decision (18-month-olds
pointed to indicate the location of a hidden toy;
12-month-olds gazed toward the location of a briefly
presented face) followed by a measure of postdecision
persistence. This measure relies on the fact that one
will wait longer for a reward if one thinks that one’s
response was correct; it has proven to be a reliable
proxy for decision confidence in rats and chimpanzees
(Beran etal., 2015; Lak etal., 2014). Infants persisted
more after making a correct compared with an incorrect
decision. Importantly, they did so in the total absence
of external feedback on their performance. Further,
their persistence did not vary with stimulus or task
properties (i.e., memorization delay for 18-month-olds;
duration of the masked-face presentation for 12-month-
olds), but rather constituted a characteristic pattern in
which confidence decreased with difficulty for correct
judgments but increased for errors (Lak etal., 2014).
This suggests that infants’ postdecision persistence does
not simply reflect properties of external events (e.g.,
that the face was presented for 100 ms vs. 300 ms), but
rather is contingent on the accuracy of their own deci-
sions and reflects a subjective evaluation of perfor-
mance (see Goupil & Kouider, 2016).
One might still argue that it remains debatable
whether postdecision persistence unambiguously
reflects second-order (i.e., metacognitive) computations
(Fleming & Daw, 2017) and that, ultimately, it may be
difficult to make conclusions about this issue on the
basis of behavioral data alone. Fortunately, a fourth
approach using neuroimaging provides more decisive
evidence, as it allows researchers to dissociate brain
areas associated with metacognitive versus cognitive
processes. Regarding confidence, experiments with
both rats and monkeys revealed that pharmacological
inactivation of specific prefrontal areas impairs meta-
cognitive sensitivity while sparing perceptual decisions
(Lak etal., 2014; Miyamoto etal., 2017). Similarly, meta-
cognition causally involves the prefrontal cortex in
human adults (Fleming & Dolan, 2012), and decision
confidence is automatically encoded in the ventrome-
dial prefrontal cortex even when no reports are required
from the participants (Lebreton, Abitbol, Daunizeau, &
Pessiglione, 2015). The metacognitive system support-
ing error monitoring also elicits characteristic neural
activity originating in the anterior cingulate cortex and
reflected in an electroencephalographic potential called
the error-related negativity (ERN). This component is
thought to reflect a second-order, postdecisional mecha-
nism signaling a mismatch between a just-made (incor-
rect) decision and the choice that should have been
made on the basis of the available evidence (Charles,
King, & Dehaene, 2014).
Consistent with the idea that core metacognition
emerges early in development, studies have shown that
some of these key prefrontal structures are already
functional in infancy (Dehaene-Lambertz & Spelke,
2015; Goupil & Kouider, 2016). Although unfortunately
no study to date has measured neural markers of deci-
sion confidence in young children, a few studies have
shown that the mechanisms of error monitoring sub-
tended by the anterior cingulate cortex are already
Developing a Reflective Mind 405
functional (Tamnes, Walhovd, Torstveit, Sells, & Fjell,
2013). For instance, we recently found that following
perceptual decisions, infants as young as 12 months of
age display an ERN component after making an incor-
rect choice (Goupil & Kouider, 2016).
Core Metacognition
The evidence reviewed above suggests that human
infants are already endowed with rudimentary forms of
metacognition. Several lines of research suggest that
these abilities primarily rely on a phylogenetically
ancient metacognitive system that is not strictly tied to
explicit reports. Indeed, as mentioned above, confidence
and error monitoring are present in several other species,
including rats and monkeys (Hampton, 2009). Further-
more, neuroimaging and computational-modeling stud-
ies in human adults have revealed that confidence and
error monitoring can be reduced to simple computa-
tional mechanisms (Fleming & Daw, 2017; Yeung &
Summerfield, 2012) and can automatically induce self-
regulation even when they are not introspectively accessed
(Logan & Crump, 2010; Nieuwenhuis, Ridderinkhof, Blom,
Band, & Kok, 2001). Consequently, modern frameworks
emphasize the fact that metacognition does not reduce to
its explicit manifestations and stress the importance of
implicit forms of self-monitoring and regulation (Proust,
2012; Shea etal., 2014).
Building on these views, we here propose that start-
ing in the first years of life, humans are already endowed
with a system of core metacognition allowing them to
automatically evaluate and regulate their own cogni-
tion. Core metacognition is complementary to other
core systems that evolved to fulfill a dedicated function
(e.g., representing objects) and on which flexible and
explicit capacities build later in development (Carey,
2009; Spelke & Kinzler, 2007). Whereas other core sys-
tems represent properties of the external world, core
metacognition specifically evolved to enable the moni-
toring of internal cognitive states, allowing endogenous
engagement in self-regulation. Specifically, core metacog-
nition encompasses any mechanism whereby a first-
order (cognitive) representation (e.g., a belief) is
assessed through a second-order (metacognitive) pro-
cess that evaluates its quality (e.g., the likelihood that
the belief is correct given the sensory evidence) without
necessarily reaching awareness or being represented
explicitly. Importantly, core metacognition is not simply
equivalent to cognitive control: It concerns processes
that use metacognitive representations (i.e., representa-
tions that contain some information about the proper-
ties of an underlying cognitive representation; e.g., the
reliability of a belief) for self-regulation (for more
details, see Shea etal., 2014; for an alternative position
on this issue, see Carruthers, 2009). Such a system
would be present not only in humans but also in other
animals whenever their neural architecture implements
second-order structures evaluating first-order represen-
tations computed in sensorimotor and other associative
As with other core systems, core metacognition can
be considered an innate component relying on the mat-
uration of dedicated brain structures that were shaped
by evolution to constrain and optimize our representa-
tional space. However, the emergence of core metacog-
nition is certainly protracted compared with other core
systems, such as naive physics. Indeed, whereas sensory
systems develop rapidly over the first year of life, the
prefrontal cortex shows a dramatic increase in synaptic
density and long-range connectivity only by the end of
the first year (Dehaene-Lambertz & Spelke, 2015) and
substantially matures until adolescence (Lebel, Walker,
Leemans, Phillips, & Beaulieu, 2008). Thus, although the
structures that support core metacognition are already
functional during the first years of life, they remain
largely immature and undergo substantial development
throughout childhood. So far, the youngest age at which
core metacognition has been observed is 12 months
(Goupil & Kouider, 2016). Whether the maturation of
the prefrontal cortex and its long-range connections are
sufficient to support metacognition in infants below 1
year of age remains an open question.
Signature Limits of Core Metacognition
If core metacognition emerges early, why does self-
reflection appear so limited and unreliable until much
later in childhood, as evidenced by decades of research
in developmental psychology? As mentioned above, a
first developmental constraint concerns the maturation
of the prefrontal areas supporting core metacognition
that undergo substantial maturation during childhood.
But beyond this slow maturation, the core metacogni-
tive system presents several intrinsic limitations. Below,
we characterize these restrictions and detail how the
core system interacts with the late-developing system
of explicit metacognition.
Core systems are defined not only by a dedicated
function but also by distinctive signature limits (e.g.,
the core system dedicated to object representations is
limited to a set size of four; Carey, 2009). A first signa-
ture limit of core metacognition is that it is restricted
to the evaluation of simple perceptual and value-based
decisions. Even human adults tend to shift to inferential,
heuristic strategies relying on the interpretation of cues
(e.g., response times)—as opposed to relying on opti-
mal metacognitive computations—to evaluate complex
decisions and reasoning (Ackerman & Thompson, 2017).
406 Goupil, Kouider
A second and perhaps more important limitation of
core metacognition is that, in and of itself, it is not suf-
ficient to enable a crucial aspect of human metacogni-
tion: the ability to share metacognitive representations
with other people (Shea etal., 2014). This latter capac-
ity relies on explicit metacognition, a heterogeneous
system that supports the exploitation of consciously
accessed metacognitive representations. Decades of
developmental research suggest that this system
requires the development of several additional abilities
(Flavell, 2000; Sodian etal., 2012) that progressively
come together to enable children to engage in explicit
metacognition and provide accurate metacognitive
First, the ability to focus on metacognitive monitor-
ing and inhibit the use of irrelevant information when
providing self-reports is limited in children. Their inac-
curate reports reflect an inability to select relevant
sources of information (e.g., error signals) while inhibit-
ing inappropriate ones (e.g., desires to succeed). This
often results in a “liberal” metacognitive bias with a
high propensity to give affirmative (e.g., “Yes I know”)
responses (i.e., overconfidence; Butterfield, Nelson, &
Peck, 1988; Lipowski, Merriman, & Dunlosky, 2013).
Importantly, this bias is also often observed in human
adults (Dunlosky & Metcalfe, 2009). Children still lack
the executive functions necessary to inhibit such pre-
potent responses and provide accurate metacognitive
reports. Indeed, executive functions, and in particular
inhibitory control, do not consistently drive behavior
before 4 years of age and continue developing until late
childhood (Posner, Rothbart, Sheese, & Voelker, 2014;
see also Roebers, 2017, for more details on this issue).
Another developmental constraint concerns the abil-
ity to shift from an implicit to an explicit mode of
processing. The latter implies directing attention toward
metacognitive representations in order to globally
broadcast them, thereby rendering them conscious and
reportable (Dehaene, Lau, & Kouider, 2017). To illus-
trate this contrast between implicit and explicit meta-
cognition, we can return to the neural signatures of
error monitoring mentioned above. Although neural
markers of implicit error monitoring are observable in
the anterior cingulate cortex following incorrect
responses and are reflected in the ERN component
regardless of introspective access to the error, conscious
error detection elicits broader activations in a fronto-
parietal network that are reflected in a later component
called the error-related positivity (Yeung & Summerfield,
2012). This pattern is similar to what is found for con-
scious versus unconscious perceptual processes in
adults and infants, as the former generally involves
early responses that are restricted in time and space
whereas the latter evokes late, long-lasting, and wide-
spread activations (Dehaene etal., 2017; Kouider etal.,
2013). Taken together, these data suggest that con-
sciously accessing metacognitive representations
involves additional neural processes entailing their
global broadcasting. Given that the mechanisms
enabling conscious access are already in place during
the first year of life (Kouider etal., 2013), it is plausible
that, even in infancy, metacognitive representations
eventually become conscious when they represent a
very strong signal (e.g., when a decision differs greatly
from the option favored by sensory evidence). Yet it is
also probable that, initially, metacognitive representa-
tions remain for the most part unconscious, either
because their representational content remains too
weak to capture attention or because they are overwrit-
ten by stronger alternative signals (e.g., external feed-
back, desires to succeed).
Finally, beyond executive functions and conscious
access, language and culture play a crucial role in build-
ing the late-developing system (Flavell, 2000; Sodian
etal., 2012). Interactions with caregivers shape and
normalize explicit aspects of metacognition (Roebers,
2017; Schneider, 2008), and cultural norms deeply mod-
ulate individual strategies when giving self-reports (Ma
etal., 2014). Learning a set of linguistic expressions
that are efficient to share metacognitive representations
in one’s own culture is a challenge, as shown by the
fact that the acquisition of cognitive-state vocabulary
lags behind the acquisition of other lexicons (Bretherton
& Beeghly, 1982) and that even in human adults, effi-
cient communication of confidence depends on linguis-
tic convergence (Fusaroli etal., 2012). Thus, whereas
core metacognition is likely to be an innate module
evolved through natural selection (like other core sys-
tems; Carey, 2009), the emergence of explicit metacog-
nition is largely constrained by the development of
higher-order cognitive functions and culturally situated
We suggest that a core system of metacognition appears
very early in development, whereas explicit and human-
specific forms of metacognition slowly emerge to progres-
sively allow children to communicate their metacognitive
representations to others. In the field of mind reading,
distinctions between implicit and explicit processes have
also been used to account for the finding that young
children’s spontaneous behaviors reveal a sensitivity to
other people’s beliefs years before they manage to pass
explicit false-belief tasks (Low & Perner, 2012), and simi-
lar arguments have been made regarding empathy
(Heyes, 2018). How are these two aspects of mind reading
and empathy related to the two metacognitive systems?
Some evidence suggests that explicit mind reading and
metacognition develop together (Flavell, 2000; Lockl &
Developing a Reflective Mind 407
Schneider, 2007), but little empirical data are available
regarding implicit aspects. A stimulating avenue for
future research will thus be to examine the relation-
ships between these constructs while considering the
fact that they cannot be reduced to their explicit mani-
festations. Another promising avenue for future research
will be to examine the potential involvement of meta-
cognition in early learning. Metacognition is a privi-
leged tool to optimally acquire new information, as it
allows organisms to assess their own knowledge states
and flexibly adapt their strategies to learn optimally in
the absence of external feedback (Guggenmos, Wilbertz,
Hebart, & Sterzer, 2016). At the earliest stages of devel-
opment, when everything remains to be learned, this
ability might be one of the key ingredients that allow
young children to learn actively and optimally.
Recommended Reading
Goupil, L., & Kouider, S. (2016). (See References). An empiri-
cal article revealing that core metacognitive mechanisms
are already present in 12- and 18-month-old infants.
Proust, J. (2012). (See References). An opinion piece discuss-
ing the possibility that metacognition does not reduce to
its explicit manifestations.
Shea, N., Boldt, A., Bang, D., Yeung, N., Heyes, C., & Frith, C. D.
(2014). (See References). An opinion article arguing that
whereas implicit metacognition is shared with several ani-
mal species and serves functions at the subpersonal level,
explicit metacognition has evolved to fulfill a social function.
Sodian, B., Thoermer, C., Kristen, S., & Perst, H. (2012). (See
References). A review on the development of metacogni-
tion in young children.
Action Editor
Randall W. Engle served as action editor for this article.
We thank J. Sackur and J. Proust for discussions on the
Declaration of Conflicting Interests
The author(s) declared that there were no conflicts of interest
with respect to the authorship or the publication of this
This research was supported by funding from the Agence Natio-
nale de la Recherche (ANR-17-EURE-0017) and from the Euro-
pean Research Council to S. Kouider (METAWARE project).
Ackerman, R., & Thompson, V. A. (2017). Meta-reasoning:
Monitoring and control of thinking and reasoning. Trends
in Cognitive Sciences, 21, 607–617.
Balcomb, F. K., & Gerken, L. (2008). Three-year-old children
can access their own memory to guide responses on a
visual matching task. Developmental Science, 11, 750–760.
Beran, M. J., Perdue, B. M., Futch, S. E., Smith, J. D., Evans, T. A.,
& Parrish, A. E. (2015). Go when you know: Chimpanzees’
confidence movements reflect their responses in a com-
puterized memory task. Cognition, 142, 236–246.
Bretherton, I., & Beeghly, M. (1982). Talking about inter-
nal states: The acquisition of an explicit theory of mind.
Developmental Psychology, 18, 906–921.
Butterfield, E. C., Nelson, T. O., & Peck, V. (1988). Developmental
aspects of the feeling of knowing. Developmental Psycho-
logy, 24, 654–663.
Carey, S. (2009). The origin of concepts. Oxford, England:
Oxford University Press.
Carruthers, P. (2009). Mindreading underlies metacognition.
Behavioral & Brain Sciences, 32, 164–182.
Charles, L., King, J.-R., & Dehaene, S. (2014). Decoding
the dynamics of action, intention, and error detection
for conscious and subliminal stimuli. The Journal of
Neuroscience, 34, 1158–1170.
Dehaene, S., Lau, H., & Kouider, S. (2017). What is conscious-
ness, and could machines have it? Science, 358, 486–492.
Dehaene-Lambertz, G., & Spelke, E. S. S. (2015). The infancy
of the human brain. Neuron, 88, 93–109.
Dunlosky, J., & Metcalfe, J. (2009). Metacognition. Thousand
Oaks, CA: SAGE.
Flavell, J. H. (2000). Development of children’s knowl-
edge about the mental world. International Journal of
Behavioral Development, 24, 15–23.
Fleming, S. M., & Daw, N. D. (2017). Self-evaluation of deci-
sion-making: A general Bayesian framework for metacog-
nitive computation. Psychological Review, 124, 91–114.
Fleming, S. M., & Dolan, R. J. (2012). The neural basis of
metacognitive ability. Philosophical Transactions of the
Royal Society B: Biological Sciences, 367, 1338–1349.
Fusaroli, R., Bahrami, B., Olsen, K., Roepstorff, A., Rees, G.,
Frith, C., & Tylén, K. (2012). Coming to terms: Quantifying
the benefits of linguistic coordination. Psychological
Science, 23, 931–939.
Ghetti, S., Hembacher, E., & Coughlin, C. A. (2013). Feeling
uncertain and acting on it during the preschool years: A
metacognitive approach. Child Development Perspectives,
7, 160–165.
Goupil, L., & Kouider, S. (2016). Behavioral and neural
indices of metacognitive sensitivity in preverbal infants.
Current Biology, 26, 3038–3045.
Goupil, L., Romand-Monnier, M., & Kouider, S. (2016).
Infants ask for help when they know they don’t know.
Proceedings of the National Academy of Sciences, USA,
113, 3492–3496.
Guggenmos, M., Wilbertz, G., Hebart, M. N., & Sterzer, P.
(2016). Mesolimbic confidence signals guide perceptual
learning in the absence of external feedback. eLife, 5,
Article e13388. doi:10.7554/eLife.13388.001
Hampton, R. R. (2009). Multiple demonstrations of metacog-
nition in nonhumans: Converging evidence or multiple
mechanisms? Comparative Cognition Behavior Reviews,
4, 17–28.
408 Goupil, Kouider
Heyes, C. (2018). Empathy is not in our genes. Neuroscience
and Biobehavioral Reviews, 95, 499–507.
Kouider, S., Stahlhut, C., Gelskov, S. V., Barbosa, L. S., Dutat,
M., de Gardelle, V., . . . Dehaene-Lambertz, G. (2013).
A neural marker of perceptual consciousness in infants.
Science, 340, 376–380.
Lak, A., Costa, G. M., Romberg, E., Koulakov, A. A., Mainen,
Z. F., & Kepecs, A. (2014). Orbitofrontal cortex is required
for optimal waiting based on decision confidence.
Neuron, 84, 190–201.
Lebel, C., Walker, L., Leemans, A., Phillips, L., & Beaulieu,
C. (2008). Microstructural maturation of the human brain
from childhood to adulthood. NeuroImage, 40, 1044–1055.
Lebreton, M., Abitbol, R., Daunizeau, J., & Pessiglione, M.
(2015). Automatic integration of confidence in the brain
valuation signal. Nature Neuroscience, 18, 1159–1167.
Lipowski, S. L., Merriman, W. E., & Dunlosky, J. (2013).
Preschoolers can make highly accurate judgments of
learning. Developmental Psychology, 49, 1505–1516.
Lockl, K., & Schneider, W. (2007). Knowledge about the mind:
Links between theory of mind and later metamemory.
Child Development, 78, 148–167.
Logan, G. D., & Crump, M. J. C. (2010). Cognitive illusions of
authorship reveal hierarchical error detection in skilled
typists. Science, 330, 683–686.
Low, J., & Perner, J. (2012). Implicit and explicit theory of
mind: State of the art. The British Journal of Developmental
Psychology, 30, 1–13.
Ma, Y., Bang, D., Wang, C., Allen, M., Frith, C., Roepstorff, A.,
& Han, S. (2014). Sociocultural patterning of neural activ-
ity during self-reflection. Social Cognitive and Affective
Neuroscience, 9, 73–80.
Miyamoto, K., Osada, T., Setsuie, R., Takeda, M., Tamura, K.,
Adachi, Y., & Miyashita, Y. (2017). Causal neural network
of metamemory for retrospection in primates. Science,
355, 188–193.
Nieuwenhuis, S., Ridderinkhof, K. R., Blom, J., Band, G. P.,
& Kok, A. (2001). Error-related brain potentials are
differentially related to awareness of response errors:
Evidence from an antisaccade task. Psychophysiology,
38, 752–760.
Posner, M. I., Rothbart, M. K., Sheese, B. E., & Voelker, P.
(2014). Developing attention: Behavioral and brain mech-
anisms. Advances in Neuroscience, 2014, Article 405094.
Proust, J. (2012). Metacognition and mindreading: One or
two functions? In M. J. Beran, J. L. Brandl, J. Perner,
& J. Proust (Eds.), Foundations of metacognition (pp.
234–251). Oxford, England: Oxford University Press.
Roebers, C. M. (2017). Executive function and metacogni-
tion: Towards a unifying framework of cognitive self-
regulation. Developmental Review, 45, 31–51.
Schneider, W. (2008). The development of metacognitive knowl-
edge in children and adolescents: Major trends and implica-
tions for education. Mind, Brain, and Education, 2, 114–121.
Shea, N., Boldt, A., Bang, D., Yeung, N., Heyes, C., & Frith,
C. D. (2014). Supra-personal cognitive control and meta-
cognition. Trends in Cognitive Sciences, 18, 186–193.
Sodian, B., Thoermer, C., Kristen, S., & Perst, H. (2012).
Metacognition in infants and young children. In M. J.
Beran, J. L. Brandl, J. Perner, & J. Proust (Eds.), Foundations
of metacognition (pp. 119–133). Oxford, England: Oxford
University Press.
Spelke, E. S., & Kinzler, K. D. (2007). Core knowledge.
Developmental Science, 10, 89–96.
Tamnes, C. K., Walhovd, K. B., Torstveit, M., Sells, V. T., &
Fjell, A. M. (2013). Performance monitoring in children
and adolescents: A review of developmental changes
in the error-related negativity and brain maturation.
Developmental Cognitive Neuroscience, 6, 1–13.
Vo, V. A., Li, R., Kornell, N., Pouget, A., & Cantlon, J. F.
(2014). Young children bet on their numerical skills:
Metacognition in the numerical domain. Psychological
Science, 25, 1712–1721.
Yeung, N., & Summerfield, C. (2012). Metacognition in
human decision-making: Confidence and error monitor-
ing. Philosophical Transactions of the Royal Society B:
Biological Sciences, 367, 1310–1321.
... Here, we offer a detailed account of how metacognition supports curiosity. A large body of research indicates that metacognition does not reduce to its conceptual, meta-representational forms, and that preverbal children and some nonhuman animals possess rudimentary metacognitive ressources (Goupil & Kouider, 2019;Proust, 2012Proust, , 2019Shea et al., 2014, see Appendix 1). We argue that intrinsic curiosity involves at least such minimal forms of metacognitive monitoring and regulation that allow agents to identify and satisfy their informational needs, which we refer to as "core" or "procedural" metacognition. ...
... Low confidence already predicts information-seeking early on in development (Coughlin, Hembacher, Lyons, & Ghetti, 2014;Goupil & Kouider, 2019;Lapidow, Killeen, & Walker, 2022). For instance, in a perceptual task, 3-to 5-year-old children preferentially seek additional informationinstead of responding by themselvesin conditions in which they also report low confidence on a picture-based scale (Coughlin et al., 2014). ...
... 3. In human ontogeny, children display behavioral markers suggesting that they experience feelings of curiosity and confidence before they can conceptually represent themselves as having these mental states (Baer & Kidd, 2022;Goupil & Kouider, 2019). 4. In human adults, dissociations have been documented in metacognitive monitoring according to task demands, between feelingbased and concept-based incompatible evaluations (Koriat & Ackerman, 2010;Nussinson & Koriat, 2008). ...
Full-text available
Curious information-seeking is known to be a key driver for learning, but characterizing this important psychological phenomenon remains a challenge. In this article, we argue that solving this challenge requires qualifying the relationships between metacognition and curiosity. The idea that curiosity is a metacognitive competence has been resisted: researchers have assumed both that young children and non-human animals can be genuinely curious, and that metacognition requires conceptual and culturally situated resources that are unavailable to young children and non-human animals. Here, we argue that this resistance is unwarranted given accumulating evidence that metacognition can be deployed procedurally, and we defend the view that curiosity is a metacognitive feeling. Our metacognitive view singles out two monitoring steps as a triggering condition for curiosity: evaluating one's own informational needs, and predicting the likelihood that explorations of the proximate environment afford significant information gains. We review empirical evidence and computational models of curiosity, and show that they fit well with this metacognitive account, while on the contrary, they remain difficult to explain by a competing account according to which curiosity is a basic attitude of questioning. Finally, we propose a new way to construe the relationships between curiosity and the human-specific communicative practice of questioning, discuss the issue of how children may learn to express their curiosity through interactions with others, and conclude by briefly exploring the implications of our proposal for educational practices.
... Though traditional accounts have long assumed that metacognition (i.e., the ability to evaluate one's own cognitive representation) is limited in children below four (e.g., Flavell, 1999), recent research outside the language domain has suggested that certain core aspects of metacognition develop much earlier, long before children can talk about their own cognition (Balcomb and Gerken, 2008;Geurten and Bastin, 2019;Ghetti et al., 2013;Goupil and Kouider, 2019). For example, there is evidence that infants are able to estimate confidence (i.e., the likelihood that a decision is correct (Kepecs et al., 2008;Pouget et al., 2016)) years before they can provide metacognitive verbal reports (Balcomb and Gerken, 2008;Geurten and Bastin, 2019;Goupil and Kouider, 2016;Hembacher and Ghetti, 2014;Kuzyk et al., 2019;Vo et al., 2014). ...
... It may be that confidence directly reflects properties of the decision-making process (e.g., the distance between accumulated evidence and a decision bound, or an evaluation of decision time) (Kiani and Shadlen, 2009;Pereira et al., 2021). Alternatively, it could be that confidence reflects core metacognitive monitoring even in young children (Goupil and Kouider, 2019). The finding that speaker idiosyncrasy can differentially impact first-look accuracy and post-decision persistence favours this latter interpretation. ...
... In sum, our work converges with a growing body of evidence suggesting that monitoring confidence is a fundamental ability that enables humans to actively and adaptively respond to their environment from a very young age (Ghetti et al., 2013;Goupil and Kouider, 2019). It extends previous results by showing that toddlers' capacity for confidence monitoring is not restricted to the evaluation of simple perceptual decisions, but extends to socially-informed conventional knowledge. ...
We studied the fundamental issue of whether children evaluate the reliability of their language interpretation, that is, their confidence in understanding words. In two experiments, 2-year-olds (Experiment 1: N = 50; Experiment 2: N = 60) saw two objects and heard one of them being named; both objects were then hidden behind screens and children were asked to look toward the named object, which was eventually revealed. When children knew the label used, they showed increased postdecision persistence after a correct compared with an incorrect anticipatory look, a marker of confidence in word comprehension (Experiment 1). When interacting with an unreliable speaker, children showed accurate word comprehension but reduced confidence in the accuracy of their own choice, indicating that children’s confidence estimates are influenced by social information (Experiment 2). Thus, by the age of 2 years, children can estimate their confidence during language comprehension, long before they can talk about their linguistic skills.
... Interns self-reflect on their strengths, weaknesses, and progress made during the semester and discuss specific areas for improvement with the supervisor. The self-reflection is a metacognitive activity (Goupil & Kouider, 2019), which provides an opportunity for interns to think aloud about their own abilities, while they also develop a plan of action for future practice. ...
Full-text available
Small special education programs (SSEPs) are composed of limited faculty tasked with educating interns dispersed across large geographical areas (Reid, 1994). These needs underscore a call for more flexible educational program options. Moreover, Kebritchi et al. (2017) found professors in higher education institutions sought a variety of instructional methods to critically respond to barriers experienced by SEPPs. The purpose of this article is to highlight virtual methods utilized by SSEPs for field experiences, modeling,coaching, feedback, supervision, and partnerships to leverage faculty expertise effectively and efficiently, to expand recruitment in programs, and to support teacher retention efforts. Using the Council for Exceptional Children (CEC) and Collaboration for Effective Educator Development, Accountability, and Reform (CEEDAR) High Leverage Practices (HLPs) of Instruction, Collaboration, and Assessment (McLeskey et al., 2017), this article will look behind the virtual lens to uncover how SSEPs faculty can support interns using a developmental and scaffolded approach.
... Still, if metacognitive processes and multisensory congruence develop in the first few years of life, the sense of perceptual reality should also become richer but also more confusing at times, as infants and children develop (see Goupil & Kouider, 2019). Similarly, the capacity to entertain counterfactual hypotheses (e.g. ...
In a Bayesian brain, every perceptual decision will take into account internal priors as well as new incoming evidence. A reality monitoring system—eventually providing the agent us with a subjective sense of reality avoids us them being confused about whether our experience is perceptual or imagined. Yet not all confusions we experience mean that we wonder wonder whether we may be imagining: some confused experiences feel clearly perceptual but still feel not right. What happens in such confused perceptions, and can the Bayesian brain explain this kind of confusion? In this paper, we offer a characterisation of perceptual confusion and argue that it requires our subjective sense of reality to be a composite of several subjective markers, including a categorical one that can clearly identify an experience as perceptual and connecting us to reality. Our composite account makes new predictions regarding the robustness, the non-linear development and the possible breakdowns of the sense of reality in perception.
Young children are often dependent on learning from others and to this effect develop heuristics to help distinguish reliable sources from unreliable sources. Where younger children rely heavily on social cues such as familiarity with a source to make this distinction, older children tend to rely more on an informant’s competence. Little is known about the cognitive mechanisms that help children to select the best informant; however, some evidence points toward mechanisms such as metacognition (thinking about thinking) and theory of mind (thinking about other’s thoughts) being involved. The goals of the current study were to (a) explore how the monitoring and control components of metacognition may predict selective social learning in preschoolers and (b) attempt to replicate a reported link between selective social learning and theory of mind. In Experiment 1, no relationship was observed across the measures. In Experiment 2, only selective social learning and belief reasoning were found to be related as well as when both experiments’ samples were combined. No links between selective social learning and metacognition were observed in the two experiments. These results suggest that theory of mind is a stronger correlate of selective learning than metacognition in young children. The implications regarding the kind of tasks used to measure metacognition are discussed.
Full-text available
It is a major challenge to identify assessment instruments for group work that not only support students’ learning processes and progress but also reflect a valid and reliable result based on individual efforts. In this chapter, we discuss the use of electronic posters (e-posters) as a multimodal assessment instrument for virtual excursions. The chapter involves a review of relevant literature on the use of posters embedded within CL. The value of posters as a multimodal assessment instrument, contributing towards the promotion of much-needed SDL skills (i.e. critical thinking, communication, and deeper learning), is central to this research. Qualitative data from students’ self assessments were analysed in order to present practical guidelines in terms of the implementation of e-posters as an assessment instrument for virtual excursions, as well as the affordances thereof in the first-year students’ learning process.
Full-text available
Virtual excursions, as an iteration of online-based learning and teaching activities, are become more popular and in this context, SDL could be regarded as an important aspect of the learning process. In this chapter, we aim to explore the overall trends and characteristics from published works on online virtual excursions in terms of SDL from the last 20 years. To this end, this chapter involves a systematic literature review conducted with a corpus of published works carefully screened based on set inclusion criteria related to virtual excursions and SDL. The process then further involved an inductive thematic and structural analysis of the publications. From this research, an overview is presented on thematic and methodological aspects of discourses on virtual excursions, but the chapter also specifically considers how engagement with SDL happens on such platforms. It is evident that the research relating to SDL regarding virtual excursions is fairly limited and that many opportunities exist for future work in this context. The chapter concludes by presenting guidelines and research lacunae, based on the reviewed publications, for setting up virtual excursions that could promote SDL as well as identifying appropriate learning strategies within this context.
Learners use certainty to guide learning. They maintain existing beliefs when certain, but seek further information when they feel uninformed. Here, we review developmental evidence that this metacognitive strategy does not require reportable processing. Uncertainty prompts nonverbal human infants and nonhuman animals to engage in strategies like seeking help, searching for additional information, or opting out. Certainty directs children’s attention and active learning strategies and provides a common metric for comparing and integrating conflicting beliefs across people. We conclude that certainty is a continuous, domain-general signal of belief quality even early in life.
This book puts forward a mechanistic account of subjective experience based on a review of the current cognitive neuroscience literature on conscious perception, attention, and metacognition. It is argued that current empirical studies are often misinterpreted. An undue focus has been placed on perceptual capacity rather than subjective experience per se. Null findings are often overemphasized despite the limited sensitivity of the methods used. A synthesis is proposed to combine the advantages and intuitions of both global and local theories of consciousness. This will be discussed in the context of our understanding of the sense of agency, emotion, rationality, culture, philosophical theories, and clinical applications. Taking insights from both physiology and current research in artificial intelligence, the resulting view directly addresses the qualitative nature of subjective experience.
In the chapter we address the remaining issues of consciousness in animals and robots. A good way to address this may be to first figure out what in principle accounts for consciousness in the human cases. From there, we can see if similar mechanisms exist or not in the animals or robots. To make this inductive generalization, we need a theory of consciousness. We introduce the perceptual reality monitoring theory, according to which some animals may not be conscious. And yet, perhaps even a robot or computer program could be. This is to say, it is a functionalist account, based on an empirical interpretation of a variant of higher-order theory of consciousness. To evaluate the theory, we should compare it against other existing theories in terms of plausibility.
Full-text available
It is well established that learning can occur without external feedback, yet normative reinforcement learning theories have difficulties explaining such instances of learning. Here, we propose that human observers are capable of generating their own feedback signals by monitoring internal decision variables. We investigated this hypothesis in a visual perceptual learning task using fMRI and confidence reports as a measure for this monitoring process. Employing a novel computational model in which learning is guided by confidence-based reinforcement signals, we found that mesolimbic brain areas encoded both anticipation and prediction error of confidence-in remarkable similarity to previous findings for external reward-based feedback. We demonstrate that the model accounts for choice and confidence reports and show that the mesolimbic confidence prediction error modulation derived through the model predicts individual learning success. These results provide a mechanistic neurobiological explanation for learning without external feedback by augmenting reinforcement models with confidence-based feedback.
Full-text available
Meta-Reasoning refers to the processes that monitor the progress of our reasoning and problem-solving activities and regulate the time and effort devoted to them. Monitoring processes are usually experienced as feelings of certainty or uncertainty about how well a process has, or will, unfold. These feelings are based on heuristic cues, which are not necessarily reliable. Nevertheless, we rely on these feelings of (un)certainty to regulate our mental effort. Most metacognitive research has focused on memorization and knowledge retrieval, with little attention paid to more complex processes, such as reasoning and problem solving. In that context, we recently developed a Meta-Reasoning framework, used here to review existing findings, consider their consequences, and frame questions for future research.
Full-text available
Are you aware how well you remember? Self-monitoring and evaluation of our own memory is a mental process called metamemory. For metamemory, we need access to information about the strength of our own memory traces. The brain structures and neural mechanisms involved in metamemory are completely unknown. Miyamoto et al. devised a test paradigm for metamemory in macaques, in which the monkeys judged their own confidence in remembering past experiences. The authors combined this approach with functional brain imaging to reveal the neural substrates of metamemory for retrospection. A specific region in the prefrontal brain was essential for meta mnemonic decision-making. Inactivation of this region caused selective impairment of metamemory, but not of memory itself. Science , this issue p. 188
Full-text available
People are often aware of their mistakes, and report levels of confidence in their choices that correlate with objective performance. These metacognitive assessments of decision quality are important for the guidance of behavior, particularly when external feedback is absent or sporadic. However, a computational framework that accounts for both confidence and error detection is lacking. In addition, accounts of dissociations between performance and metacognition have often relied on ad hoc assumptions, precluding a unified account of intact and impaired self-evaluation. Here we present a general Bayesian framework in which self-evaluation is cast as a “second-order” inference on a coupled but distinct decision system, computationally equivalent to inferring the performance of another actor. Second-order computation may ensue whenever there is a separation between internal states supporting decisions and confidence estimates over space and/or time. We contrast second-order computation against simpler first-order models in which the same internal state supports both decisions and confidence estimates. Through simulations we show that second-order computation provides a unified account of different types of self-evaluation often considered in separate literatures, such as confidence and error detection, and generates novel predictions about the contribution of one’s own actions to metacognitive judgments. In addition, the model provides insight into why subjects’ metacognition may sometimes be better or worse than task performance. We suggest that second-order computation may underpin self-evaluative judgments across a range of domains.
Full-text available
Humans adapt their behavior not only by observing the consequences of their actions but also by internally monitoring their performance. This capacity, termed metacognitive sensitivity [1, 2], has traditionally been denied to young children because they have poor capacities in verbally reporting their own mental states [3–5]. Yet, these observations might reflect children’s limited capacities for explicit self-reports, rather than limitations in metacognition per se. Indeed, metacognitive sensitivity has been shown to reflect simple computational mechanisms [1, 6–8], and can be found in various non-verbal species [7–10]. Thus, it might be that this faculty is present early in development, although it would be discernible through implicit behaviors and neural indices rather than explicit self-reports. Here, by relying on such non-verbal indices, we show that 12- and 18-month-old infants internally monitor the accuracy of their own decisions. At the behavioral level, infants showed increased persistence in their initial choice after making a correct as compared to an incorrect response, evidencing an appropriate evaluation of decision confidence. Moreover, infants were able to use decision confidence adaptively to either confirm their initial choice or change their mind. At the neural level, we found that a well-established electrophysiological signature of error monitoring in adults, the error-related negativity, is similarly elicited when infants make an incorrect choice. Hence, although explicit forms of metacognition mature later during childhood, infants already estimate decision confidence, monitor their errors, and use these metacognitive evaluations to regulate subsequent behavior.
In academic and public life empathy is seen as a fundamental force of morality – a psychological phenomenon, rooted in biology, with profound effects in law, policy, and international relations. But the roots of empathy are not as firm as we like to think. The matching mechanism that distinguishes empathy from compassion, envy, schadenfreude, and sadism is a product of learning. Here I present a dual system model that distinguishes Empathy¹, an automatic process that catches the feelings of others, from Empathy², controlled processes that interpret those feelings. Research with animals, infants, adults and robots suggests that the mechanism of Empathy¹, emotional contagion, is constructed in the course of development through social interaction. Learned Matching implies that empathy is both agile and fragile. It can be enhanced and redirected by novel experience, and broken by social change.
The controversial question of whether machines may ever be conscious must be based on a careful consideration of how consciousness arises in the only physical system that undoubtedly possesses it: the human brain. We suggest that the word “consciousness” conflates two different types of information-processing computations in the brain: the selection of information for global broadcasting, thus making it flexibly available for computation and report (C1, consciousness in the first sense), and the self-monitoring of those computations, leading to a subjective sense of certainty or error (C2, consciousness in the second sense). We argue that despite their recent successes, current machines are still mostly implementing computations that reflect unconscious processing (C0) in the human brain. We review the psychological and neural science of unconscious (C0) and conscious computations (C1 and C2) and outline how they may inspire novel machine architectures.
Executive function and metacognition are higher-order cognitive processes that undergo steady improvements throughout childhood. They are highly relevant to daily functioning in various domains, including academic achievement. Both concepts have been intensively researched, but surprisingly little literature has sought to connect them theoretically and empirically. In the present review, I elaborate on the similarities between these concepts from a developmental perspective, including the definitions, developmental timetables, factors that lead to changes over time, and relations to academic achievement and intelligence. Simultaneously, the differences between these two domains of cognitive development are discussed. These include, in particular, the relative neglect of quantifying monitoring within research on executive functions and the disregard for the neuropsychological underpinnings of metacognition. Finally, this paper presents several avenues for future research and proposes a possible unifying framework of cognitive self-regulation that integrates executive function and metacognition and may lead to a better understanding of the emergence of cognitive self-regulation in development.