Skip to navigation | Skip to content

Implicit and Explicit Theory of Mind

[Editor’s note: This is a slightly revised version of a talk given by Rita Astuti for the 2014 European Society for Philosophy and Psychology Annual Conference as part of a panel on “Theory of Mind and the Social Mind”. As Astuti explains, her presentation to philosophers and psychologists on this occasion can be seen as a companion piece to a presentation on the same topic given to an audience of anthropologists on an earlier occasion, and that has already been published in AOTC.]

A couple of years ago I gave a talk to an audience of anthropologists about a topic that is referred to in the psychological and philosophical literatures as Theory of Mind ( Briefly, Theory of Mind refers to the human ability to attribute such things as desires, knowledge, ignorance, (true or false) beliefs, etc., to other people and to interpret the visible behaviour of these people in terms of these invisible mental states.

Among the anthropologists listening to that previous talk, there were a few who work in societies of the Pacific where people are said to subscribe to the so-called “opacity of mind doctrine” (see Rumsey & Robbins 2008). This doctrine stipulates that one cannot know what is inside other people’s minds, and that it is in fact morally wrong to impute mental states to others. Doing so amounts to a kind of assault on people’s mental privacy and on their personal integrity.

Anthropologists working in these societies have documented various practices that follow on from this doctrine, most especially as they pertain to the socialization of children. For example, adults never attribute intentions to an infant’s babble, but regard these vocalisations as no more than natural reflexes (Ochs 1982: 91-92); and if older children are caught reading any meaning into the babble of their younger siblings, they are scolded and told that the infant is doing adadadada rather than saying adadadada (Schieffelin 2008: 433-434). Along the same lines, children are taught not to say that such and such a person was sad or that she was hungry, but to report that she cried or that she ate a lot. And to avoid being accused of gossiping – a very serious moral breach – they learn to report people’s direct speech, rather than their (allegedly unknowable) thoughts or opinions (Schieffelin 2008: 436).

The opacity of mind doctrine is a wonderfully clear example of a culturally specific and highly developed folk theory of the human mind and in my talk I used this example to make one key point to my fellow anthropologists: that we should not be tempted into using folk theories of the mind, such as the opacity of mind doctrine, to make claims about cross-cultural variation in people’s Theory of Mind. This is because people’s folk theories are explicit reflections about the human mind, whereas Theory of Mind – as understood by psychologists – operates implicitly as individuals, automatically and unconsciously, follow somebody’s gaze, interpret her utterances, or predict her next move on the basis of her desires, emotions or beliefs.

To put it bluntly, I was warning my anthropological colleagues not to be naïve about the reach of this kind of evidence: culturally specific doctrines and folk theorising. This kind of evidence is adequate to tackle the level of explicit discourse and normative practices, but it is inadequate to make claims about the implicit workings of human cognition.

What I want to do today is to make a different but complementary point, better suited to this particular audience: namely, that philosophers and psychologists who work on Theory of Mind should not be naïve about the reach of their evidence into what they also call “the explicit.”

I should add straight away that I’m not trying to pick a fight with philosophers and psychologists. My ultimate aim, in the spirit of the Human Mind Project,[1] is to highlight the need for genuine inter-disciplinary collaboration, which can only happen if the differences between each constituency are fully recognised, including the relative strengths and weaknesses of their approaches.


Let me start with a bit of background.

For years, research on Theory of Mind has been dominated by one single and very robust finding: that children under the age of 4 fail the (in)famous False Belief Task. This task seeks to measure whether children understand that other people might have (false) beliefs that do not correspond to the way the world actually is. This is a significant measure of Theory of Mind development because the fact that a child is able to predict a person’s action (that Sally will look for a chocolate bar in the place where she left it) by way of representing her false belief (that the chocolate bar is still where she left it, even though it actually isn’t), truly proves that the child is reasoning about the person’s mental states (Sally’s belief about where the chocolate is), rather than about the actual state of the world (in fact, the chocolate bar has been surreptitiously moved in Sally’s absence).

But then, after hundreds of studies that, again and again, reported the very same finding – that 3 year-olds fail in this reasoning exercise – a new generation of False Belief Tasks started to show that infants as young as 13 months were able to pass the test (see Baillargeon, Scott & He 2010 for a review). Given that these infants are pre-linguistic, the tasks used to test them had to be non-verbal, by contrast with the tasks that have typically been used with toddlers. Thus, so called non-elicited response False Belief Tasks (used with infants) use indirect measures, such as the infant’s anticipatory looking, while elicited response False Belief Tasks (used with toddlers) require the child to answer a direct question (He, Bolz & Baillargeon 2011). To illustrate: the research subject is presented with a character, be it a caterpillar or a doll or an actor, who places a desired object in one location and then leaves the scene; in the character’s absence, the object is moved to a different location, at which point the character re-enters the scene. In non-elicited response tasks, the experimenter will track the infant’s gaze to see which location she expects the character to move towards: the one where the character falsely believes the object to be or the one where the object actually is; in elicited-response tasks, the toddler is asked where the character will look for the object.[2]

Given this difference in experimental design, it has been suggested that the puzzling discrepancy between infants’ and toddlers’ performance can be explained in light of the distinction between implicit and explicit Theory of Mind. I’m well aware that the distinction is not used by everyone: certainly not by those psychologists (e.g., Baillargeon and her colleagues), who stress the continuity of Theory of Mind competence across development, and thus attribute children’s failure on elicited response tasks to external constraints that have nothing to do with their core knowledge of mental states (for example, lack of inhibition: once a toddler knows where the chocolate really is, she will find it difficult to inhibit her knowledge and allow herself to say that Sally will look for it where it is not). For those who, by contrast, stress the qualitative difference between infants’ and older children’s competence, the distinction between implicit and explicit Theory of Mind works to highlight the limitations of infants’ abilities (who pass non-elicited response tasks) by comparison with those of the older children (who pass the elicited response tasks).

For example, in a recent paper on the cultural evolution of Theory of Mind, Cecilia Heyes and Chris Frith (2014) frame the distinction between implicit and explicit Theory of Mind around the analogy between ‘script’ and ‘mind reading’ which, they suggest, should be used more productively than it has been done so far. The analogy was originally used to highlight that both script and mind reading are based on extracting meaning from signs: a mark on paper means “dog” and a stretching of an arm means “she wants that glass.” But equally if not more significant is the fact that both script and mind reading rely on specialised neurocognitive mechanisms and that both can be impaired by genetically inherited disorders. These two characteristics might suggest that both types of reading have evolved by genetic means. We know, however, that this is not the case for print reading, which is a very recent historical invention that gets transmitted through cultural means. And if this is the case for print reading – with its neural specialisation and associated genetic disorders – why not for mind reading as well? (see Strickland & Jacob 2015 for a robust critique and Heyes & Frith’s response).

This argument paves the way for a dual approach to Theory of Mind, which distinguishes between mind reading abilities – those of the implicit variety – that are supported by cognitive mechanisms that have evolved (and are inherited) genetically, and mind reading abilities – those of the explicit variety – that are supported by cognitive mechanisms that have evolved (and are inherited) culturally.

Implicit mind reading, according to Heyes and Frith, is what infants can do, as they automatically and efficiently track mental states. It is also what people with autism cannot do, because of genetic impairments in the neurocognitive mechanisms that have evolved for this purpose. Conversely, explicit mind reading is what toddlers need to learn in order to be able to deliberate and to talk about mental states – something they must  do in order to “pass” elicited-response versions of the False Belief task; it is what people with autism can learn to do, through extensive training and tuition; and it is what normally developing adults cannot do when demands are made on their executive functions – demands that, by contrast, do not disrupt their implicit mind reading abilities.


So, what is my problem with the psychologists’ use of “the explicit” in these debates?

First, I should make it clear that I do not have a problem with Heyes and Frith’s headline argument: that explicit mind reading is inherited culturally and that, as a result, it is shaped by culture. As they themselves note, this must be true for most, if not all human neurocognitive skills, given the fundamentally cultural nature of our species; as anthropologist Clifford Geertz pithily put it, without culture any one of us would be a mental basket case (1973: 49).[3]

My problem is rather with the way in which “the explicit”, in the hands of the psychologists, gets cut down to size: specifically, to the size of a few notable experimental tasks.


Let me use a couple of examples to illustrate what I mean.

The first one is the famous “natural experiment” involving the emergence of Nicaraguan Sign Language. For peculiar historical reasons, there was no sign language in Nicaragua until the late 70s. In due time, there were two cohorts of signers: one older but with a more impoverished language, and one younger but with a far richer language. In a language elicitation task, where participants had to respond to short video clips that revealed the protagonist’s state of ignorance or her frustrated desires, the first cohort used far fewer (or none at all) mental state terms than the second cohort and, unlike the second cohort, they spectacularly failed the False Belief Task. A few years later, having spent more time with the second cohort and having enriched their language as a result, the first cohort improved significantly on both counts.

These are fascinating results, with deep implications about the link between language and Theory of Mind. However, it seems to me that the conclusions drawn from these studies should be interrogated with a qualitatively different kind of evidence.  Thus, when I read the claim that “adults who had no congenital cognitive deficits, but whose language was incomplete, failed to fully understand the beliefs of others” and that they “struggled to understand others’ mental states” (Pyers & Senghas 2009: 810), I want to ask: How do these individuals behave in their daily lives? How do they take care of their children? Do they ever tell them white lies? What do they do when their children tell them lies? How do they react if they find out that their spouse is cheating on them? What do they do if they want to cheat on their spouse? And so on…

It is possible that the psychologists who have been working for years with these Nicaraguan individuals and their families have this kind of rich observational data and that the reason this kind of data does not find its way into their published work is that psychology journals do not make space for it (this is indeed my experience when publishing in psychological venues: my ethnographic data is often deemed too detailed and anecdotal). But if this is the case, I would suggest that this is an editorial problem, with serious theoretical implications, which needs some very urgent attention!


Let me now turn to another example: a study conducted by psychologists Andreas Mayer and Birgit Traüble in one of the Pacific islands that make up the state of Independent Samoa. Samoans are one of the populations that subscribe to the opacity of mind doctrine I mentioned at the outset. Accordingly, anthropologists have reported that Samoans ‘‘act as if they could not know or do not care about what is in someone else’s mind’’ (Duranti 2006: 33); that they display a “cultural dispreference for talking about or making claims about what another might feel, what another might think, or what another is about to do” (Ochs & Schieffelin 1984: 290); and that, in the context of legal disputes, whether someone has committed a misdeed intentionally or not is deemed irrelevant (Shore 1982: 182). To the Samoans, what matters are the effects of what people do and say, not whether they meant to do it or say it – indeed, it has been reported that the sentence “I didn’t mean it” is literally impossible in the Samoan language (Duranti 1993: 33).

So, this is the cultural context in which the 300 or so children tested by Mayer and Traüble grow up. The children, aged between 3 and 14, were given a False Belief Task that asked them to predict where one of their school mates would look for a wooden toy. This is how the study was set up (I go into some details here because, as we shall see, the details are important):

Two children entered the testing room and one of them was asked to hide the toy under one of three cups; this child was then sent out and told to wait outside.

The remaining child was invited to play a trick on the absent class mate by moving the toy under a different cup. When this was done, she was asked where the absent class mate would look for the toy when she came back.

After offering her False Belief answer, the child was sent back to the class-room.

At this point, the child who had been sent out returned to the testing room to find the toy on the table.

Another child was called in, was asked to hide the toy, was sent out and was told to wait outside.

The child who had previously been sent out was now invited to play a trick on the absent class mate and was asked the False Belief question.

And so on…

The results show a very different developmental trajectory to the one observed in Euro-America: while Samoan children did get better with age, the majority still did not succeed in the False Belief Task before they were 8, not to mention the fact that one-third of the 10–12-year-olds still failed.

It must be stressed that the authors of this study do not rush to the conclusion that it is the opacity of mind doctrine that explains the children’s performance. Specifically, they note that while these children grow up surrounded by “the doctrine” and the practices that follow from it, they also grow up in large extended families, have large numbers of siblings, and live in “open houses” where social life is complex and intense. In light of this, as the authors rightly point out, one might predict a precocious rather than a delayed development of mind reading abilities.

So, what are we to make of these results? What should our next move be? My suggestion is that if we really want to grasp what Samoan “explicit” theory of mind is like, when and how it develops, we need to leave the testing room and follow the children as they move and talk and listen and play with other people (rather than in experimental isolation); we need to follow them in the classrooms, in the playgrounds, in the open houses, in the kitchens, on the beaches and up the mango trees…

I’m not suggesting that what happens in the testing room is insignificant; but the testing room is just one – very weird – context, and what happens in it needs to be interpreted in light of what happens elsewhere in the lives of the children (this, of course, applies just as much to Euro-American children who, like their Samoan counterparts, also live rich social lives). Indeed, the problem is that the children themselves cannot but interpret the False Belief Task through the lenses of their social experience elsewhere: games that involve hiding and deception; strategies for competing over food or other resources;[4] stories that they read at school or hear on the radio or in church, which might involve tricksters and their moral downfall or brilliant successes, and so on.

This might all sound far-fetched, but the point is simple: it is only if we manage to imagine the testing situation from the view-point of the child – a child who comes to it with a variety of rich experiences that she has shared with other people – that we can hope to interpret what her performance really means. Let me illustrate the point with an anecdote from my field site in Madagascar – an anecdote which might actually provide some useful insights into what might have happened in the Samoan testing room.

A few years back, I tried out the False Belief tasks with a few children in the village where I have been carrying out my research for many years now.[5]

I used a similar design to the one described above: I first identified the child to be tested – in this particular case a shy 4 year old boy – and I then recruited a second child – an older brother – to help out with the procedure. I got both children to witness the hiding of a coin under one of two coconut shells; I then asked the older boy to run off on a random errand: a common thing for a parent or older sibling to expect from a child, but quite unusual for me, the resident anthropologist, since this was not my own adoptive family.

Once the older child was out of sight, I got the test child to join me in moving the coin from one coconut shell to the other. I then asked the standard False Belief question: when the older child returns, where will he look for the coin? On this occasion, the younger child failed the False Belief Task, although I was impressed by the fact that he actually pointed to anything at all. At this young age, Malagasy children often use a very effective strategy when directly addressed by an adult: a prolonged silence, accompanied by the elaborate twisting of a corner of their shorts or sarongs!

Anyway, the insight came after the test boy failed the task. His older sibling promptly returned with the water I had asked him to fetch. Thinking that I might use the test situation to boost the development of the 4 year old’s Theory of Mind, I asked the older boy to look for the coin. To my surprise, he went straight for the coconut where, in his absence, we had moved the coin to! As he grabbed the coin and ran off, he shouted: “that’s why you sent me to fetch the water!”

This older child – he was probably nine or ten – was clearly one step ahead of the game! He interpreted the testing situation in terms of his real life experience: Who was I to order him to fetch water? Surely, the reason I was sending him away was to give his little brother the chance to win the coin…

His sophisticated and, for sure, “explicit” Theory of Mind raises the question of how he would have performed in the False Belief Task: might he have “failed” the task by answering that, like he did himself, the child who had been sent to fetch water would choose to look under the “wrong” coconut? If so, clearly it would be the task, not the child, that failed.

In the light of this, I’m wondering whether there was something of the same going on in Samoa, at least among the older children. It’s not hard to imagine that, as they were made to wait outside the testing room, they might have asked themselves what on earth was going on inside that they were not allowed to participate in. They had just been asked to hide a toy under one of three cups, so…

If this was their experience – perhaps modelled on other experiences with hiding games – they might have later used it to figure out that the next child who was to be sent out of the room (whose false belief they had to diagnose) would have had the exact same thoughts and would have come to the exact same conclusion: that in her absence a deceptive trick was being played on her![6]

The general point I want to draw from this is that, as cultural psychologist Michael Cole never tires of stressing, no testing procedure is ever culture-free (see Cole 1996, 1999). This doesn’t just mean that we have to be careful not to “scare away” the psychological competence that we are testing for by using ecologically inappropriate techniques – which is undoubtedly a valid concern; it also means that we need to understand the specific ways in which our techniques are being understood and experienced by those we are testing. And the point, of course, is that how we interpret our results – for example, as evidence of failure or of “super-success” on the False Belief task – entirely depends on our being aware of these understandings and experiences.


So, how do we ensure that we know what is going on? I have hinted at the answer already, and it’s important to stress that it is not an easy answer with quickly deliverable solutions. The answer is that we need to understand the social life of the children we study, and we need to understand it in its full complexity. For example, we need to understand the historical specificity of their experiences of family life: where siblings, rather than parents, might be the care providers; where silence might be an acceptable response for children who are deemed to have no wisdom; where, for the same reason (i.e., lack of wisdom), adults might use absurd lies to soothe their children (as I have observed in my field site in Madagascar); or where adults might pinch or slap their kids hard and challenge them not to cry in order to “toughen them up” and teach them how to withstand punishment (Stafford 1995: 51-52); and so on. We need to understand what schooling represents in the lives of the children: a place where, as Michael Cole puts it, children who, at the market, are ever so smart, seemingly become ever so dumb (1996: 74); a place where one identifies as an “individual” instead of identifying as a member of a kinship group; a place where for the first time one becomes a “citizen”; and so on.

Above all – and here I’m ashamedly shifting into preaching mode – we (that is: you!) need to understand that “culture” is not a “thing” that people have or to which they belong; a thing which people hold “explicitly” as they subscribe to certain norms and beliefs; a thing that makes people adopt particular attitudes and orientations and which, as “input”, might affect their cognitive development and overall cognition. Culture is not something that can be categorised as being of one type (for example, individualistic) or another (for example, collectivistic); and it is not something that can be used as an independent variable to study the development of Theory of Mind.

If it sounds like I care (a lot), let me explain that the whole of the anthropological tradition – at least of the British and social variety – has worked hard to dislodge this particular view of culture as a coherent, pre-packaged entity that is “explicitly” inscribed in people’s beliefs and which shape their worldview. Rather, British social anthropology has pursued the study of how people produce and transmit cultural artefacts – ranging from ways of walking and eating, to ways of raising children, to ritual incantations, fishing techniques, folk theories of the mind and what not – in the context of dynamic processes of social and historical transformation. This is why British social anthropologists have turned their backs on the idea that one can capture “culture” by asking people what they do, what they believe, which norms they follow.

Instead, we adopt a method tellingly called “participant observation”: that is, we participate in all aspects of the social life of a particular place and, while we participate, we also observe. In practice, what this means is that we get up in the morning and we join the flow of social life, doing things with other people: going fishing, going to the market, attending funerals, visiting new born babies, gossiping, cooking and eating, playing dominos and so on. And then, at some convenient time, we retire to write our field notes, which describe and analyse what we have participated in.

The essential ingredient of this methodology is time: time to learn the language so that we can work without intermediaries and fully participate in the social life with other people; time to transition from clumsiness to fluency, not just in spoken language, but also in our posture, in our walking and sitting, in our washing and eating; time to learn the local etiquette, and what is funny, and what is worrying; time to learn and experience first-hand what makes people tick, their beliefs and expectations about themselves and others; and time to notice that the very same people can be recklessly “individualistic” in one context of social life (say, as they trade fish in the market) and staunchly “collectivistic” in another (say, when they finance the construction of an ancestral tomb).

As you can well imagine, participant observation does not yield systematic data sets. But what it does yield is an understanding “from within” of the highly textured and context-dependent experience of being a person – a child, a parent, a teacher, a deaf signer – in Samoa, in Madagascar, in Nicaragua, and so on. It is only on the basis of this understanding, I would argue, that experimental tasks and the peculiar social dynamic that they introduce can be properly contextualised, and the results they produce accurately interpreted.

And when this is all done, we – the anthropologists – might even tempt you – the philosophers and psychologists – to join us “in the wild” in order to take the full measure of people’s “explicit” Theory of Mind.



Baillargeon, R., Scott, R. M., & He, Z. (2010). False-belief understanding in infants. Trends in Cognitive Sciences, 14, 110–118.

Bloch, M. (2006) L’anthropologie cognitive à l’épreuve du terrain : l’exemple de la théorie de l’esprit. Leçons inaugurale du Collège de France, n°184, delivered 23 February 2006.

Cole, M. (1999).  Culture-free versus culture-based measures of cognition. In R. J. Sternberg (Ed.), The nature of cognition, pp. 645-664. MA: The MIT Press.

Duranti, A. (1993). Intentions, Self, and Responsibility: An Essay in Samoan Ethnopragmatics, in J.H. Hill & J.T. Irvine (Eds.), Responsibility and Evidence in Oral Discourse, pp. 24–47. Cambridge: Cambridge University Press.

Duranti, A. (2006). The social ontology of intentions. Discourse Studies, 8, 31–40.

Duranti, A. (2008). Further reflections on reading other minds. Anthropological Quarterly, 81 (2), 483-494.

He, Z., Bolz, M. & Baillargeon, R. (2011) 2.5-year-olds succeed at a verbal anticipatory-looking false-belief task. British Journal of Developmental Psychology, 30, 14–29.

Mayer, A. & Traüble, B. E. (2012). Synchrony in the onset of mental state understanding across cultures? A study among children in Samoa. International Journal of Behavioral Development, 37(1), 21–28

Ochs, E. (1982). Talking to children in Western Samoa. Language in Society, 11, 77-104.

Ochs, E. & Schieffelin, B. (1984). Language acquisition and socialization: Three developmental stories. In R. Shweder & R. LeVine, Culture theory: Mind, self, and emotion, pp.276-320. Cambridge: Cambridge University Press.

Pyers, J. E., & Senghas, A. (2009). Language promotes false-belief understanding: Evidence from learners of a new sign language. Psychological Science, 20 (7), 805–812.

Rubio-Fernández, P. & Geurts, B. ((2013). How to pass the false-belief task before your 4th  birthday. Psychological Science, 24, 27-33.

Rumsey, A. & Robbins, J. (2008). Social Thought and Commentary Section: Anthropology and the Opacity of Other Minds, Anthropological Quarterly, 81 (2), 407-494.

Schieffelin, B. B. (2008). Speaking Only Your Own Mind: Reflections on Talk, Gossip and Intentionality in Bosavi (PNG). Anthropological Quarterly, 81 (2), 431-44.

Stafford, C. (1995). The Roads of Chinese Childhood: Learning and Identification in Angang. Cambridge: Cambridge University Press.

Strickland, B. & Jacob, P. (2015). Why reading minds is not like reading words. Blog of The International Cognition and Culture Institute: [accessed 24 April 2015].

Tomasello, M. & Rakoczy, H. (2003). What makes human cognition unique? From individual to shared to collective intentionality, Mind & Language, 18 (2), 121-147.

  1. [1]The Human Mind Project was the organiser of the conference panel. Not to be confused with the Human Brain Project, the Human Mind Project includes the contribution of the arts and humanities into the study of human nature and more generally calls for an integrated and interdisciplinary approach to the study of the human mind.
  2. [2]But see Rubio-Fernández & Geurts 2013 for a simplified verbal False Belief Task, which 3-year olds have been able to pass.
  3. [3]In a similar vein, although from a rather different theoretical perspective, Tomasello and Rakoczy note that “if we imagine a human child born on to a desert island, somehow magically kept alive by itself until adulthood, it is possible that this adult’s cognitive skills would not differ very much – perhaps a little, but not very much –from those of other great apes” (2013: 121).
  4. [4]Duranti (2008: 489-490) reports a short interaction between a mother, and two siblings, one 6 and the other 9 years old: the interaction is about food and about the 9 year old’s frustration at the younger child’s request for the biggest piece of banana. The significant point noted by Duranti is that the younger child does not actually ask for the bigger piece of banana; it is the older child who infers the mental state of her younger sibling from her (unreasonable) actions.
  5. [5]This was inspired by Maurice Bloch, who at the time was also doing fieldwork in Madagascar and was using the False Belief Task – which he administered to children – to trigger discussions among, and elicit reflections from, adults about the mind, about knowledge, about lying (see Bloch 2006).
  6. [6]Since this study involved three cups, one might argue that, had the children reasoned along these lines, they should have randomly selected one of the two cups that, from the perspective of the child who had been sent out, had been empty and thus equal candidates for hiding the toy. This, however, would have required quite a sophisticated understanding of chance, which is probably unlikely at this age.

Please join our mailing list to receive notification of new issues