By: Anthony Aubel
One of the central topics in philosophy of cognitive science is the Theory of mind, or mindreading, which refers to the ability to infer and predict others’ mental states. The study of mindreading is divided by two competing theories that dominate the field: Theory-Theory (TT) and Simulation Theory (ST). Proponents of TT argue that our understanding of others’ mental states involves employing an information-rich set of folk psychological (i.e. common sense) knowledge that can explain and predict others behaviors and mental states. In contrast, ST supporters claim that we understand others by simulating their mental states; that is, we replicate their experiences within ourselves. By adopting pretend beliefs and pretend desires that we think the other person has, we simulate these pretend mental states to explain and predict the other person’s behavior (Davies and Stone 1995b). In the field of neuroscience, the discovery of so-called mirror neurons by Vittorio Gallese, Giacomo Rizzolatti, and their team of researchers from University of Parma, Italy (i.e. Parma group), has fueled interests in the study of mindreading. Mirror neurons are observed to fire (activate) in specific brain regions when an individual performs an action and also when that individual observes someone else performing a similar action. These findings have led researchers to propose mirror neuron activity as a possible neural basis (i.e. neural correlate) for mindreading.
In this essay, I plan to explore the relation between mirror neurons and competing theories of mindreading particularly with respect to TT and ST. The central question I will attempt to address is to what extent, if any, the activity of mirror neurons offer compelling evidence in favor of either TT or ST. Drawing from the debates between Shannon Spaulding and Vittorio Gallese et al., I will defend Spaulding’s position, arguing that mirror neuron activity does not offer convincing evidence in support of ST, whereas it lends stronger, albeit limited, support to the TT account of mindreading. Based on this analysis, I conclude that mirror neurons provide only a partial picture of cognitive processes involved in mindreading and hence they cannot be considered the sole basis for either theory.
Background: Mirror neurons and mindreading:
To understand how the connection between mirror neurons and mindreading theories came about, let’s take a closer look at their discovery and function. In the early 1990s, a group of neurophysiologists at the University of Parma led by Giacomo Rizzolatti found that certain neurons in a region of macaque monkeys’ brain (called area F5) that are responsible for controlling hand and mouth movements become active both when the monkey performed a specific action and when it observed another individual monkey or human performing the same action (Rizzolatti et al., 1996). Most neurons in this area tend to fire during motor actions such as grasping, holding, or manipulating objects. This observation sparked the proposal that mirror neurons provide the neural basis (i.e. correlate) for understanding actions, imitations, and social cognition (Gallese, Keysers, & Rizzolatti, 2004; Rizzolatti et al., 2001; Rizzolatti & Craighero, 2004). The discovery led to further research to explore if such mirror neuron systems also exist in humans, and if yes, whether or not they play a similar role in social cognition. Subsequent research by the Parma group showed that mirror neurons are in fact involved in a range of cognitive processes, including action understanding, imitation, empathy, and language processing (Rizzolatti & Sinigaglia, 2010). Vittorio Gallese and Alvin Goldman in particular have been the strongest proponents of the proposal that mirror neurons are at the core of human social cognition and play a central role in our ability to understand and predict the mental states of others (Gallese et al., 2004). Such claims have generated significant interest and debates within philosophy of cognitive science and neuroscience circles as to the extent and significance of the potential link between neurons and mindreading. Both TT and ST supporters have attempted to use these findings and incorporate them into their respective theories, leading to a heated debate around the significance of mirror neurons in understanding others' mental states and behavior.
Theories of mindreading: TT and ST:
Let us now consider a more in-depth review of the two theories of mindreading; TT and ST. As mentioned earlier, mindreading (or the theory of mind), refers to our ability to attribute mental states, such as beliefs, intentions, and desires to oneself as well as others in order to explain and predict behavior. In the philosophy of cognitive science literature, TT and ST have been the two dominating yet competing theories that attempt to explain this ability. TT proposes that humans in particular possess an innate (or learned) ‘Theory of Mind (ToM)’ that uses folk psychological concepts allowing us to understand and predict the mental states and/or behaviors of others. TT entails a set of general principles and causal relations that rely on information-rich processes enabling individuals to make inferences about the mental states of others based on their behavior (Gopnik & Wellman, 1992). Simulation theorists, in contrast, argue that individuals understand and predict the mental states of others by simulating or re-creating those states by using our own minds (Goldman, 2006). ST is often stated as putting oneself in another’s shoes, so to speak, and using one’s own cognitive mechanisms and experiences to generate a form of pretend beliefs of another person’s mental state that allows for an intuitive understanding of their thoughts and experiences. Hence, ST is sometimes considered to be an information-poor mindreading process.
Proponents from both camps, TT and ST, have used the evidence from studies on mirror neurons mentioned earlier as a basis to claim that mirror neuron activity provides evidence in favor of their respective positions. For instance, Gallese et al., have argued that mirror neurons provide a clear neural mechanism in favor of simulation process. They claim that the evidence suggests that mindreading is supported by the observation of activity patterns of mirror neurons that strongly correlate with instances where other’s actions are observed, allowing individuals to recreate the mental states within their own minds. This, they claim, is direct evidence for the ST account and is sufficient to enable mindreading. On the other hand, TT proponents acknowledge the importance of mirror neurons in social cognition, but they remain skeptical about the extent of mirror neuron’s role in the mindreading process; they do not consider it to be the sole or primary mechanism for mindreading. Instead, they see mirror neurons as one of many cognitive processes and mechanisms that contribute to mindreading, along with other rich sources of information, such as background knowledge about people and social structures (Spaulding, 2012). This latter position is the view that I intend to support in the remainder of this essay.
Evidence and critical onsiderations:
Let us take a closer look now at the evidence and claims made by both TT and ST about the role of mirror neurons in mindreading in order to evaluate how they hold in critical considerations.
In her paper Mirror neurons are not evidence for the Simulation Theory, Shannon Spaulding highlights several reasons why mirror neuron activity does not provide sufficient evidence to support ST. She provides a compelling set of arguments that questions the claims about the causal link between mirror neuron activity and the process of simulation in mindreading. For instance, Spaulding contends that in ST, one of the requirements is that the observer’s mental state serves as a type of representational surrogate for the target individual’s mental state. However, she points out that mirror neuron activation does not necessarily give rise to such surrogate representations since the observer’s mental states are often genuine or attenuated (as Goldman and Gallese suggest) rather than a pretend or simulated one (Goldman, 2006). Proponents of ST, such as Goldman and Gallese, who strongly support the view that mirror neurons provide evidence for ST, are quick to assert that similarity between the observer’s and target’s mirror neuron activity automatically entails simulation. But this transitive claim is not justified because merely being in a similar cognitive state does not establish that simulation has occurred. There is a distinction made between a mere duplication of neural states and the process of simulation. Since ST suggests that we use our own cognitive resources to recreate the target’s experiences in our minds, then it is possible for the observer to attain the same cognitive state through other cognitive means; those other means could very well be that of TT type.
The position that Spaulding argues for is also consistent with evidence supporting that mirror neurons are insufficient by themselves to determine mental states. For instance, certain facial expressions or other single observed behavior can be associated with multiple mental states. Psychological studies have shown, for example, that facial reactions of blushing can signify a multitude of states such as anger, joy, embarrassment, or just a reactionary hot flash (Jackson, Meltzoff and Decety, 2005). These findings indicate that additional contextual information combined with other cognitive processes may be necessary for accurate determination of the underlying mental state. Mirror neuron activity, therefore, by itself cannot solely account for the complexity of such observations and this insufficiency undermines the strong claims of ST that mirror neurons are the primary mechanism for mindreading in simulation processes.
Spaulding also presents another compelling piece of evidence that runs against the sufficiency of mirror neuron activity in support of ST; that is the lack of a clear distinction between simulation that is considered offline versus online. Nichols and Stich (2003, pp. 132–135) argue that genuine mental simulation requires that the cognitive process be taken “offline” and provided with non-standard inputs (Spaulding, 2012). This distinction is of critical importance for ST proponents. Specifically, they argue that ‘online’ simulation is what is being used when one is predicting others’ actions while ‘offline’ simulation is applied when one attempts to understand and attribute mental states to another. But there is ambiguity in the data as to whether the activation of mirror neurons corresponds to offline simulation or online simulation or both for that matter. This inability to differentiate between offline and online simulation implies that mirror neuron activity is not exclusive to mental state attribution or action prediction.
Gallese and Goldman may object to this argument by claiming that the lack of clear offline and online distinction is not a fatal flaw for ST and suggest further research may still identify more specific patterns of mirror neuron activity that correlate with offline simulation. This could very well be possible that future research reveals a more clear distinction between offline and online simulation. But the current state of empirical evidence, as it stands, does not support the ST’s reliance on mirror neuron activity as the primary mechanism for mindreading. In fact, there is sufficient evidence from psychological studies in the areas of attention and associative learning that can explain mirror neuron activation by means of these alternative cognitive processes.
Moreover, several studies show that mirror neurons can be activated during tasks that do not necessarily involve mindreading, such as the execution phase of actions and observers’ response to inanimate objects. This evidence runs contrary to the argument for the exclusive role of simulation attributed to mirror neuron activity. These findings may not directly negate the role of mirror neurons in simulation. A case could be made that these alternative explanations could still be compatible with ST and contribute to the overall process of simulation. However, this would not address the central issue that is being raised here; that mirror neuron activity alone is not sufficient evidence for ST. If such cognitive processes as mentioned earlier (i.e. attentional or associative learning) seem to contribute to mirror neuron activation, then the claim that mirror neurons are exclusively responsible for simulation in mindreading is certainly weakened.
Several neuroscientific studies (Brass et.al, 2007; Calder et. al., 2000) on involvement of multiple brain regions in mindreading provide evidence that further undermine the ST’s claim on mirror neuron activity as primary mechanism for mindreading. Studies indicate that the process of mindreading involves a wide and diverse range of cognitive processes that use various parts of the brain regions that extend beyond the mirror neuron system. For instance, brain regions known as superior temporal sulcus, the medial prefrontal cortex, and temporoparietal junction have all been implicated in the mindreading process. Such evidence also raises questions about the modularity of mindreading mechanisms. Mental modularity refers to the idea that the organization of the mental structures have, at least partly, distinct functions that supervene on the underlying neural structure. Proponents of ST often argue for a specific and dedicated simulation module. But this is in contrast with the evidence for multiple brain regions being involved in the mindreading process and it would lead to oversimplification in our explanations. Even if mirror neurons were involved in a broader networks highlighted above to facilitate simulation, the evidence still supports a more complex and integrated model.
Let us now examine below how such evidence may in fact provide some limited support for TT, and not necessarily ST.
The extent of evidence in support of TT:
Interestingly, the evidence cited above from the involvement of multiple brain regions seems to be more closely aligned with TT’s position, which emphasizes the integration of various cognitive processes in mindreading. However, several of the main proponents of ST that strongly support the claim about mirror neurons being the hallmark of mindreading are led to conclude no relation to TT. Spaulding provides the following quote that captures Gallese and Goldman’s view with reference to their position on TT:
“In the first article linking mirror neurons and ST, Vittorio Gallese and Alvin Goldman argue, “The point is that [mirror neuron] activity is not mere theoretical inference. It creates in the observer a state that matches that of the target. This is how it resembles the simulation heuristic. Nothing about TT leads us to expect this kind of matching” (Gallese and Goldman 1998, p. 498). Both ST proponents and theorists studying mirror neurons have argued that mirror neurons are strong evidence in favor of ST over TT (my emphasis) (Gallese and Goldman 1998; Goldman 2006, 2009; Gordon 2005; Hurley 2005; Iacoboni 2009).”
According to TT, however, mindreading relies on the integration of multiple sources of information (i.e. information-rich aspect), none of which can really be considered as necessary or sufficient on their own. In line with this evidence, the role of mirror neurons is better suited to TT account since it acknowledges their contribution as part of a larger, more complex system. In order to construct and elaborate folk psychological models of other people, this integrated perspective is indeed necessary and critical for understanding and predicting behavior.
Gallese, in his paper Motor abstraction: a neuroscientific account of how action goals and intentions are mapped and understood, provides an interesting proposal that draws a link between the properties of mirror neuron system and the so-called embodied simulation. By embodied simulation, Gallese means to say that our cognitive processes that are involved in the simulation process of understanding others’ mental states are grounded in the activation of sensory-motor systems in the brain. He posits that this type of embodied simulation describes the mirror neuron system’s functional mechanism. He claims that this link enables “...pre-linguistic forms of action and intention understanding” (Gallese, 2009). This claim may seem consistent with ST’s account for the role mental states play in linguistic structures in mindreading through embodied cognition and could be used to argue that an understanding of such linguistic structures is grounded in the same neural simulations that underlie mirror neuron activity.
While embodied cognition offers an interesting perspective on the role of language in ST, it fails to address information-rich contents that form the basis of the knowledge and prior experiences in understanding others’ minds. It does not adequately capture how we use knowledge of social norms, individual differences in personalities, and other abstract or symbolic concepts that play an important role in making inferences about others' mental states. The latter characteristics are essential part of the mindreading process and often relies on information-rich, contextual properties. This is more in accord with TT type explanation than ST.
While the evidence discussed does seem to provide stronger support for TT account of mindreading than it is for ST, it is important to recognize that the support for TT is still limited. Mirror neurons provide only a partial picture of cognitive processes involved in mindreading and hence they cannot be considered the sole basis for either theory. This is because rather than providing direct support for TT, some of the evidence mainly demonstrates that mirror neuron activity is not sufficient to fully account for mindreading. So even though TT may account for more factors than ST, it may still not capture the entire range of cognitive mechanisms and processes that contribute to mindreading.
In this essay, we explored the extent to which the evidence for mirror neuron activity supports either TT or ST in mindreading. The analysis of debates between Spaulding and Gallese et al. reveals that mirror neurons provide limited support for TT but not for ST, as they represent only a part of the complex cognitive processes involved in understanding the mental states of others. This underscores the need for considering various cognitive processes and a range of contextual factors in mindreading research, rather than exclusively relying on mirror neurons as a basis for either theory.
Reference List
Brass, M. et al. (2007) ‘Investigating Action Understanding: Inferential Processes versus Action Simulation’, Current Biology, 17(24), pp. 2117–2121. doi:10.1016/j.cub.2007.11.057.
Davies, M. and Stone, T. (1995) Mental simulation: evaluations and applications-reading in mind and language. John Wiley & Sons.
Gallese, V. (1998) ‘Mirror neurons and the simulation theory of mind-reading’, Trends in Cognitive Sciences, 2(12), pp. 493–501. doi:10.1016/s1364-6613(98)01262-5.
Gallese, V. (2007) ‘Before and below ‘theory of mind’: embodied simulation and the neural correlates of social cognition’, Philosophical Transactions of the Royal Society B: Biological Sciences, 362(1480), pp. 659–669. doi:10.1098/rstb.2006.2002.
Gallese, V. (2009) ‘Mirror Neurons, Embodied Simulation, and the Neural Basis of Social Identification’, Psychoanalytic Dialogues, 19(5), pp. 519–536. doi:10.1080/10481880903231910.
GALLESE, V., KEYSERS, C. and RIZZOLATTI, G. (2004) ‘A unifying view of the basis of social cognition’, Trends in Cognitive Sciences, 8(9), pp. 396–403. doi:10.1016/j.tics.2004.07.002.
Goldman, A.I. (2000) ‘Folk Psychology and Mental Concepts’, ProtoSociology, 14, pp. 4–25. doi:10.5840/protosociology2000141.
GOLDMAN, A.I. (2009) ‘Mirroring, Simulating and Mindreading’, Mind & Language, 24(2), pp. 235–252. doi:10.1111/j.1468-0017.2008.01361.x.
Goldman, A.I. and others (2006) Simulating minds: The philosophy, psychology, and neuroscience of mindreading. Oxford University Press on Demand.
Gopnik, A. and Wellman, H.M. (1992) ‘Why the child's theory of mind really is a theory’.
GORDON, R.M. (1986) ‘Folk Psychology as Simulation’, Mind & Language, 1(2), pp. 158–171. doi:10.1111/j.1468-0017.1986.tb00324.x.
Jackson, P.L., Meltzoff, A.N. and Decety, J. (2005) ‘How do we perceive the pain of others? A window into the neural processes involved in empathy’, NeuroImage, 24(3), pp. 771–779. doi:10.1016/j.neuroimage.2004.09.006.
Nichols, S. and Stich, S.P. (2003) ‘Mindreading: An Integrated Account of Pretence, Self-Awareness, And: An Integrated Account of Pretence, Self-Awareness, and Understanding Other Minds’.
Rizzolatti, G. and Craighero, L. (2004) ‘THE MIRROR-NEURON SYSTEM’, Annual Review of Neuroscience, 27(1), pp. 169–192. doi:10.1146/annurev.neuro.27.070203.144230.
Rizzolatti, G. and Sinigaglia, C. (2010) ‘The functional role of the parieto-frontal mirror circuit: interpretations and misinterpretations’, Nature Reviews Neuroscience, 11(4), pp. 264–274. doi:10.1038/nrn2805.
SPAULDING, S. (2010) ‘Embodied Cognition and Mindreading’, Mind & Language, 25(1), pp. 119–140. doi:10.1111/j.1468-0017.2009.01383.x.
Spaulding, S. (2012) ‘Mirror neurons are not evidence for the Simulation Theory’, Synthese, 189(3), pp. 515–534. doi:10.1007/s11229-012-0086-y.
Spaulding, S. (2013) ‘Mirror Neurons and Social Cognition’, Mind & Language, 28(2), pp. 233–257. doi:10.1111/mila.12017.