A Philosophical Perspective on the Infant Mindreading Puzzle 

Dr John Michael—Postdoctoral Fellow in the Department for Cognitive Science—Central European University


‘I know a small child who replied when asked “Are you learn­ing the viol­in”? with a           deni­al: ‘No, I already know how to play the viol­in; I’m learn­ing how to play it bet­ter’            (Millikan 2000: 55).

Over the past 35 years or so, there has been a con­tinu­ous and intense focus with­in the cog­nit­ive sci­ences upon the nature of our every­day, com­mon­sense psy­cho­logy. How is it that we can effort­lessly pre­dict and under­stand oth­er people’s beha­vi­or in most every­day situ­ations? If I see my col­league hid­ing behind a statue of Gustav Mahler while our mutu­al boss walks by on the way to a meet­ing that we should all be going to, I under­stand that she does not want to go to the meet­ing, that she does not want our boss to see her because our boss might then order her to go the meet­ing, and that she believes that hid­ing behind a statue will pre­vent our boss from becom­ing aware of her pres­ence because our boss can­not see through statues. The tac­tic makes sense to me. What this case illus­trates is that we often make sense of oth­ers’ beha­vi­or by exer­cising an abil­ity to ascribe men­tal states to them, such as beliefs, desires and inten­tions. This capa­city has vari­ously been dubbed ‘the­ory of mind’, ‘men­tal­iz­ing’, or ‘mindread­ing’ (I will use the lat­ter term here, as it is cur­rently the most prevalent).

One gen­er­al strategy that has been pur­sued in invest­ig­at­ing the nature of mindread­ing has been to com­pare dif­fer­ent kinds of agents with dif­fer­ent mindread­ing abil­it­ies – i.e. dif­fer­ent spe­cies and also human chil­dren at dif­fer­ent stages of devel­op­ment – and to attempt to cor­rel­ate these dif­fer­ences with oth­er cog­nit­ive and/or social dif­fer­ences in those agents. The task that has most com­monly been used in assess­ing wheth­er an agent is cap­able of mindread­ing has been the so-called false belief task. In a stand­ard verbal false belief task, a child (or any oth­er test sub­ject) observes as an agent places an object in loc­a­tion A and then tem­por­ar­ily departs, whereupon a second agent arrives on the scene and trans­fers the object to loc­a­tion B. When the first agent returns, the child is asked where s/he is likely to search for the object. The cor­rect answer, of course, is that the agent is likely to search in loc­a­tion A, because that is where s/he falsely believes the object to be loc­ated (Wimmer and Perner, 1983; Baron-Cohen et al., 1985). This task – first pro­posed by the philo­soph­er Daniel Dennett (Dennett 1978) – has been regarded as a lit­mus test for the capa­city to rep­res­ent the beliefs of oth­er agents because the child can’t use her own know­ledge con­cern­ing the loc­a­tion of the object to pre­dict where the agent will search. Rather, the child must dis­tin­guish the agent’s belief about the loc­a­tion of the object from her own know­ledge, and gen­er­ate a pre­dic­tion about where the agent is going to search on the basis of this belief.

For many years, it has been a robust find­ing that chil­dren under the age of four years tend to fail this task (Wellman et al. 2001), which has motiv­ated the wide­spread view that chil­dren young­er than about four don’t rep­res­ent beliefs. And yet, more recently, non-verbal stud­ies employ­ing vari­ous impli­cit meas­ures have now pro­duced a wealth of evid­ence that infants are sens­it­ive to oth­ers’ false beliefs soon after their first birth­days (Onishi and Baillargeon 2005; Buttelammn et al. 2009; Southgate et al. 2010) or even by the middle of their first year (Kovacs et al 2011; Southgate et al. 2014; for an over­view, see Christensen and Michael 2015).

The debate about how to account for this puzz­ling pat­tern of find­ings (i.e. the ‘infant mindread­ing puzzle’) has been struc­tured by a con­trast between rich and lean accounts. Rich accounts (e.g. Baillargeon et al., 2010; Kovacs et al. 2011) main­tain that infants rep­res­ent oth­ers’ beliefs by around one year or earli­er, and then offer some explan­a­tion to account for the lag in per­form­ance on expli­cit verbal false belief tasks, while lean accounts (Perner and Ruffman 2005; Zawidzki 2013) deny that chil­dren rep­res­ent beliefs before about four, and offer defla­tion­ary explan­a­tions of the infant data that do not impute the abil­ity to rep­res­ent beliefs to infants. Midway between these two extremes, Apperly and Butterfill (2009; see also Butterfill and Apperly 2013) have pro­posed a two-systems account of mindread­ing, which explains the infant data by pos­it­ing an early-emerging sys­tem (i.e. sys­tem 1) that enables infants and young chil­dren to track and reas­on about some belief-like cog­nit­ive states of oth­er agents, but not beliefs per se (until around the age of four, when the more soph­ist­ic­ated sys­tem 2 is in place).

Arguably, all three of these the­or­et­ic­al options are intern­ally con­sist­ent, inde­pend­ently motiv­ated by sens­ible the­or­et­ic­al con­sid­er­a­tions, and con­sist­ent with at least most of the exist­ing data. How, then, should we adju­dic­ate among them? In situ­ations such as this, philo­sophy can be of use in identi­fy­ing the­or­et­ic­al com­mit­ments which impli­citly shape a debate or a meth­od­o­logy, in mak­ing the terms of the debate clear, and in open­ing up over­looked altern­at­ives. In this vein, I would pro­pose tak­ing a fresh philo­soph­ic­al look at the infant mindread­ing debate against the back­ground of philo­soph­ic­al con­cep­tions of men­tal con­tent.

Specifically, I would like to point out that this debate has been and con­tin­ues to be shaped by an impli­cit com­mit­ment to a descript­iv­ist con­cep­tion of men­tal con­tent. The basic idea under­ly­ing descriptiv­ism is that the con­tents of a men­tal rep­res­ent­a­tion are indi­vidu­ated by the set of descrip­tions with which that rep­res­ent­a­tion is asso­ci­ated, and which under­pin the role that the rep­res­ent­a­tion plays in cat­egor­iz­a­tion and in infer­en­tial reas­on­ing (Harman 1987; Block 1987; Frege 1892; Russell 1905). So, for example, a rep­res­ent­a­tion may count as a rep­res­ent­a­tion of foxes if it is asso­ci­ated with a set of descrip­tions that are sat­is­fied by foxes and which thereby enable the bear­er of the rep­res­ent­a­tion to cat­egor­izes foxes cor­rectly and to draw appro­pri­ate infer­ences about them (e.g. they are mam­mals, they have four legs, etc.).

To see how the cur­rent debate is shaped by a com­mit­ment to descriptiv­ism, let us con­sider Butterfill and Apperly’s strategy (2013; cf. Apperly and Butterfill 2009). They pro­pose that it should be pos­sible to cor­rob­or­ate a the­ory about the cog­nit­ive mech­an­isms under­ly­ing an abil­ity (such as infants’ abil­ity to suc­ceed at impli­cit false belief tasks) by find­ing evid­ence of par­tic­u­lar lim­its on that abil­ity that one would expect on the basis of the the­ory in ques­tion (but which one should not oth­er­wise expect). Thus, their the­ory about the cog­nit­ive mech­an­isms under­ly­ing infants’ mindread­ing abil­it­ies entails that infants’ mindread­ing abil­it­ies should be sub­ject to the fol­low­ing sig­na­ture lim­its: for example, infants should be unable to attrib­ute beliefs to an agent when it would require the agent to draw infer­ences involving iden­tity rela­tions, an inab­il­ity to track beliefs that involve quan­ti­fi­ers (e.g. the belief that there is no object at a loc­a­tion), an inab­il­ity to per­form level‑2 perspective-taking (i.e. to take modes of present­a­tion into account), and a lack of under­stand­ing of the inter­ac­tions between beliefs and oth­er psy­cho­lo­gic­al states in pro­du­cing actions. In con­trast, if one endorses a the­ory accord­ing to which infants rep­res­ent beliefs per se, accord­ing to Butterfill and Apperly, one should not expect their mindread­ing abil­it­ies to exhib­it these sig­na­ture limits.

For example, ima­gine that an infant knows that some agent has observed that an object O is cur­rently at a loc­a­tion L, that the infant also knows that the agent has wit­nessed evid­ence that O is identic­al with O¢, but that the infant fails to infer that the agent believes that O¢ is cur­rently at L. In such a case, we would have evid­ence that the infant lacked some core know­ledge about how beliefs com­bine with each oth­er to form oth­er beliefs and/or to guide action. Butterfill and Apperly go fur­ther, though, and sub­mit that in such a case we would have evid­ence that the infant were not able to rep­res­ent beliefs at all. The under­ly­ing thought here is that the con­tent of a rep­res­ent­a­tion is fixed by the know­ledge that is asso­ci­ated with that rep­res­ent­a­tion (i.e. the know­ledge about how beliefs com­bine with oth­er men­tal states and thereby con­trib­ute to form oth­er beliefs and/or to guide action). In oth­er words, rather than con­clude that infants make faulty infer­ences about beliefs in some cases, Butterfill and Apperly con­clude that infants do not make infer­ences about beliefs at all.

Thus, it appears that the research strategy pro­posed by Butterfill and Apperly depends upon a pri­or decision to link the judg­ment as to wheth­er chil­dren are able to rep­res­ent beliefs at a par­tic­u­lar age to the ques­tion of how well chil­dren under­stand beliefs at that age (I believe that the oth­er the­or­et­ic­al altern­at­ives are sim­il­arly shaped by descriptiv­ism, although I do not have the space to argue for this claim here). Although there is noth­ing wrong with this decision, it is import­ant to real­ize that it is a non-mandatory decision, to which there are legit­im­ate alternatives.

For example, altern­at­ive con­cep­tions of men­tal con­tent with­in the broad class of causal-historical views do not appeal to the know­ledge asso­ci­ated with rep­res­ent­a­tions in order to determ­ine their con­tents. Instead, what is decis­ive on such accounts is that the rep­res­ent­a­tion in ques­tion stand, or have in the past stood, in a par­tic­u­lar kind of caus­al rela­tion with the object, prop­erty, rela­tion or kind that is its con­tent. It is import­ant to note that causal-historical accounts dif­fer sub­stan­tially in how they spe­cify the par­tic­u­lar kind of caus­al rela­tion that determ­ines con­tent, and also with respect to the fur­ther con­di­tions they intro­duce. According to one par­tic­u­lar causal-historical the­ory, namely Ruth Millikan’s (1984; 2000; 2013) tele­ose­mant­ic approach, what is decis­ive is that the cur­rent exist­ence of a rep­res­ent­a­tion be explained by the fact that it, at some point in phylo- or onto­geny, had the func­tion of co-varying with, and thereby provid­ing inform­a­tion about, its con­tent. As a res­ult, Millikan’s approach does not demand that the rep­res­ent­a­tion cur­rently stand in any kind of caus­al rela­tion with its con­tent what­so­ever, and need not ever have stood in a reli­able caus­al rela­tion – as long as it co-varied with it often enough for some con­sumer sys­tem in the organ­ism to be able to bene­fit from using it as a source of inform­a­tion about that content.

When the infant mindread­ing debate is approached from a causal-historical per­spect­ive, and more spe­cific­ally from a tele­ose­mant­ic per­spect­ive, the ques­tion of wheth­er infants and young chil­dren rep­res­ent oth­ers’ beliefs does not hinge upon their pos­sess­ing or deploy­ing some par­tic­u­lar know­ledge about beliefs. Rather, it hinges upon their hav­ing rep­res­ent­a­tion­al vehicles that have the func­tion of co-varying with oth­er agents’ beliefs, and thereby serve as a source of inform­a­tion about oth­ers’ beliefs. Thus, infants and young chil­dren may be said to rep­res­ent beliefs even though they fail to draw infer­ences about oth­ers’ beha­vi­or that one would draw as an adult who fully mas­ters the concept of belief. As a res­ult, causal-historical the­or­ies – in par­tic­u­lar of the tele­ose­mant­ic sort – make it pos­sible to take the infant mindread­ing data at face value, and accord­ingly to con­clude that infants have the abil­ity to rep­res­ent and reas­on about beliefs by as early as six months, and yet at the same time to acknow­ledge that this abil­ity under­goes sub­stan­tial fur­ther con­struc­tion through­out child­hood. One way of under­stand­ing this point would be to note that abil­it­ies can be improved without thereby becom­ing dif­fer­ent abil­it­ies, as is illus­trated by Ruth Millikan’s charm­ing anec­dote (quoted above) about the child learn­ing to play the viol­in bet­ter.




Adams, F. and Aizawa, K., (2010). “Causal Theories of Mental Content”, The Stanford Encyclopedia of Philosophy (Spring 2010 Edition), Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/spr2010/entries/content-causal/>.

Apperly, I. A., & Butterfill, S. A. (2009). Do humans have two sys­tems to track beliefs and belief-like states? Psychological Review, 116(4), 953–970. doi:10.1037/a0016923

Baillargeon, R. (1998). Infants’ under­stand­ing of the phys­ic­al world. In M. Sabourin, F. Craik, & M. Robert (Eds.), Advances in psy­cho­lo­gic­al sci­ence, Vol. 2:  Biological and cog­nit­ive aspects (pp. 503–529). Hove,  England: Psychology Press/Erlbaum (UK) Taylor & Francis.

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

Block, N. 1987, Functional Role and Truth Conditions, Proceedings of the Aristotelian      Society LXI, 1987, 157–181

Buttelmann, D., Carpenter, M., & Tomasello, M. (2009). Eighteen-month-old infants show false belief under­stand­ing in an act­ive help­ing paradigm. Cognition, 112(2), 337–342. doi:10.1016/j.cognition.2009.05.006

Butterfill, S. A., & Apperly, I. A. (2013). How to Construct a Minimal Theory of Mind. Mind & Language, 28(5), 606–637. doi:10.1111/mila.12036.

Carlson, S. M. & Moses L. J. (2001). Individuals dif­fer­ences in inhib­it­ory con­trol and children’s the­ory of mind. Child Development, 72, 1032–1053.

Call, J., & Tomasello, M. (2008). Does the chim­pan­zee have a the­ory of mind? 30 years later. Trends in Cognitive Sciences, 12(5), 187–192. doi:10.1016/j.tics.2008.02.010

Carey, S. (2009). The Origin of Concepts. Oxford University Press.

Christensen, W. & Michael, J. (2015) From two sys­tems to a multi-systems archi­tec­ture for mindread­ing. New Ideas in Psychology. Accepted on 1 January 2015. DOI: http://dx.doi.org/10.1016/j.newideapsych.2015.01.003

Clements, W. A., & Perner, J. (1994). Implicit under­stand­ing of belief. Cognitive Development, 9(4), 377–395. doi:10.1016/0885–2014(94)90012–4

Dennett, Daniel C. ‘Beliefs about Beliefs’ Behavioral and Brain Sciences 1 (1978): 568–70.

Dretske, F. (1981). Knowledge and the Flow of Information. Cambridge, MA: MIT Press.

— (1988). Explaining Behavior: Reasons in a World of Causes. Cambridge: The MIT Press.

Frege, G. 1892. Über Sinn und Bedeutung. Zeitschrift für Philosophie und philo­soph­is­che Kritik, C: 25–50.

Harman, G. (1987). (Non-solipsistic) Conceptual Role Semantics. In New Directions in Semantics, ed. Lepore, E. London: Academic Press.

Kovács, Á. M., Téglás, E., & Endress, A. D. (2010). The Social Sense: Susceptibility to Others’ Beliefs in Human Infants and Adults. Science, 330 (6012), 1830–1834. doi:10.1126/science.1190792

Millikan, R. (1984). Language, Thought, and Other Biological Categories. Cambridge: MIT Press.

— (2000). On Clear and Confused Ideas: An Essay about Substance Concepts. Cambridge: Cambridge University Press.

— (2013). An Epistemology for Phenomenology? In: Richard Brown (ed.), Consciousness Inside and Out: Phenomenology, Neuroscience, and the Nature of Experience, Springer’s series Studies in Brain and Mind.

Onishi, K. H., & Baillargeon, R. (2005). Do 15-Month-Old Infants Understand False Beliefs? Science, 308(5719), 255–258. doi:10.1126/science.1107621

Papineau, D. (1984). Representation and Explanation, Philosophy of Science 51: 550–72.

Perner, J., and T. Ruffman. (2005). Infants’ Insight into the Mind: How Deep? Science 308:212–14.

Russell, B. 1905. On Denoting. Mind 14:479–93.

Scott, R. M., Baillargeon, R., Song, H., & Leslie, A. M. (2010). Attributing false beliefs about non-obvious prop­er­ties at 18 months. Cognitive Psychology, 61(4), 366–395. doi:10.1016/j.cogpsych.2010.09.001

Southgate, V., Chavalier, C., & Csibra, G. (2010). Seventeen-month-olds appeal to false beliefs to inter­pret oth­ers’ ref­er­en­tial com­mu­nic­a­tion. Developmental Science, 13, 907–912.

Southgate, V., Johnson, M. H., Osborne, T., & Csibra, G. (2009). Predictive motor activ­a­tion dur­ing action obser­va­tion in human infants. Biology Letters, 5(6), 769–772.

Southgate, V., Johnson, M., El Karoui, I., & Csibra, G. (2010). Motor sys­tem activ­a­tion reveals infants’ online pre­dic­tion of oth­ers’ goals. Psychological Science, 21, 355–359.

Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory-of- mind          devel­op­ment: The truth about false belief. Child Development, 72(3), 655–684

Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and con­strain­ing func­tion of wrong beliefs in young children’s under­stand­ing of decep­tion. Cognition, 13(1), 103–128. doi:10.1016/0010–0277(83)90004–5

Zawidzki, T. (2013). Mindshaping, Cambridge: MIT Press.

One thought on “A Philosophical Perspective on the Infant Mindreading Puzzle ”

Comments are closed.