* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Title
Embodied cognitive science wikipedia , lookup
Animal consciousness wikipedia , lookup
Development of the nervous system wikipedia , lookup
Caridoid escape reaction wikipedia , lookup
Neural coding wikipedia , lookup
Neural oscillation wikipedia , lookup
Affective neuroscience wikipedia , lookup
Biological neuron model wikipedia , lookup
Functionalism (philosophy of mind) wikipedia , lookup
Central pattern generator wikipedia , lookup
Emotion perception wikipedia , lookup
Abnormal psychology wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Metastability in the brain wikipedia , lookup
Neuroanatomy wikipedia , lookup
Circumventricular organs wikipedia , lookup
Optogenetics wikipedia , lookup
Feature detection (nervous system) wikipedia , lookup
Mental image wikipedia , lookup
Pre-Bötzinger complex wikipedia , lookup
Synaptic gating wikipedia , lookup
Premovement neuronal activity wikipedia , lookup
Nervous system network models wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
DO MIRROR NEURONS SUPPORT SIMULATIONBASED, LOW-LEVEL MIND-READING? Title: §0. INTRODUCTION A group of neurophysiologists from Parma have discovered mirror neurons – these neurons, first discovered in the F5 ventral premotor cortex of the macaque monkey, respond both when the monkey performs certain actions and when the monkey observes someone else (monkey or human) performing those actions.1 The same researchers also discovered mirror neurons in humans.2 One neuroscientist, not from the Parma group, predicts that “mirror neurons will do for psychology what DNA did for biology” and further singles out mirror neurons as the most important discovery of the past decade.3 This is because mirror neurons may help explain what Baron-Cohen has dubbed “mind-reading,” the ability to attribute mental states to oneself and others.4 For example, some of the Parma neurophysiologists suggest that mirror neuron activity underlies our ability to learn by imitation5 and our ability to empathize with others.6 Is there a neurological basis for our mind-reading abilities? And, if so, what are the philosophical ramifications? In philosophy, there have been two main approaches to mind-reading: simulationtheory (ST) and theory-theory (TT). Consider the following example.7 Mr. Crane arrives at the airport 30 minutes late and is told that his flight left on time. Mr. Tees arrives at the airport 30 minutes late, is told that his flight was delayed, and that he missed it by only 5 See Di Pellegrino et al., Rizzolatti et al. (1988), and Rizzolatti et al. (1996a). See Fadiga et al. and Rizzolatti et al. (1996b). 3 Ramachandran, <http://edge.org/documents/archive/edge69.html>. 4 See Chapter 1 Mindblindness and Mindreading, 1-8. 5 See Rizzolatti and Craighero. 6 See Gallese. 7 See Kahneman and Tvesky, 201-208. 1 2 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? minutes. Ninety-six percent of people tested said that Mr. Tees is more upset. How do we account for this attribution of Mr. Crane’s and Mr. Tees’s mental states? According to ST, we put ourselves in Mr. Crane’s and Mr. Tees’s shoes, so to speak, and imagine how upset we would be in each of their situations. According to TT, we deploy a theory, with some folk-psychological laws about what makes travelers upset, to infer whether Mr. Crane or Mr. Tees is more upset. Alvin Goldman is one of the key proponents of ST. Since their discovery, he has suggested that mirror neurons support simulation-based mind-reading. He initially took on a modest position: enthusiasm about the philosophical prospects of mirror-neurons with respect to mind-reading. With Vittorio Gallese, he claimed: [Mirror] neurons represent a primitive version, or possibly a precursor in phylogeny, of a simulation heuristic that might underlie mindreading.8 Most recently, Goldman has argued that mirror neurons support simulation-based, low-level mind-reading (but not necessarily high-level).9 Low-level cognitive processes are unlike highlevel ones in that they occur non-explicitly. So contrary to the mind-reading of the Mr. Crane and Mr. Tees example, low-level mind-reading occurs automatically, below the conscious level. In this paper, I argue, contra Goldman, that mirror neurons do not support simulation-based, low-level mind-reading. In the first section, I explain the distinction between TT and ST and determine where Goldman’s view is located. In the second section, I explain his case for a simulation-based mirroring model of low-level mind-reading. In the third section, while I concede that Goldman paints a plausible picture of a mirroring model, I argue that he fails to show that there is something distinctively simulational about it. If 8 9 Gallese and Goldman, 498. See Goldman (2006). CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 1 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? anything, I suggest that the mirroring model has distinctively theoretical features. I therefore conclude that the discovery of mirror neurons does not favor simulational over theoretical low-level mind-reading. §1. THE DISTINCTION BETWEEN TT AND ST Goldman states: The core difference between TT and ST, in our view, is that TT depicts the mind-reading as a thoroughly ‘detached’ theoretical activity, whereas ST depicts mind-reading as incorporating an attempt to replicate, mimic, or impersonate the mental life of the target agent.10 What is this ‘detached’ theoretical activity? And what qualifies as an attempt to replicate, mimic, or impersonate? Goldman explains this distinction diagrammatically. In the following figures, ellipses represent beliefs, squares represent desires, double-circles represent decisions, hexagons represent decision-making mechanisms, diamonds represent theoretical reasoning mechanisms, and shading indicates that the mental state is pretend.11 Figure A In Figure A, we have an agent undergoing a normal decision-making procedure. She has a desire to achieve a goal (g) and a belief that some course of action (m) will achieve that goal. She inputs this desire and belief into her decision-making mechanism, which outputs her decision to perform m. 10 11 Gallese and Goldman, 497. These figures are from Gallese and Goldman, 498, and (reproduced in) Goldman (2006), 26-29. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 2 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? Figure B In Figure B, we have a TT account of mind-reading. The attributor believes that the target has a desire to achieve a goal (g) and a belief that some course of action (m) will achieve that goal. She also believes some folk-psychological law about how the target will make his decision. She then inputs these three components into her theoretical reasoning mechanism, which outputs the target’s decision to perform m. Figure C In Figure C, we have a ST account of mind-reading. The attributor believes that the target has a desire to achieve a goal (g) and a belief that some course of action (m) will achieve that goal. But she does not make use of some folk-psychological law about how the target will make his decision. Rather, she creates a pretend desire in herself to achieve a goal (g) and a CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 3 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? pretend belief that some course of action (m) will achieve that goal. She inputs these pretend components into her (hypothesized use of a) decision-making mechanism, which outputs the pretend (target’s) decision to perform m. This process simulates the target’s decision-making process for the attributor, which allows her to attribute the decision to the target. Goldman clarifies this distinction by making note of the negative and positive approaches to ST. On the negative approach, any suggestion that mind-reading is theoretical threatens ST. So ST traditionally denies that the attributor makes theoretical inferences. That is, she does not make inferences from any folk-psychological law when mind-reading. Additionally, Goldman claims that two features are central to the positive approach to ST: (1) the hypothesized role of pretend states and (2) the hypothesized use of mechanisms or processes of the same kind employed by the target.12 We can account for these two features in the above figures. (1) For example, the shaded areas in Figure C are absent in Figure B. The attributor makes use of pretend states when she creates in herself a desire to achieve a goal (g) and a belief that some course of action (m) will achieve that goal. These inputs produce a pretend output, the pretend (target’s) decision to perform m. (2) Further, the hexagon in Figure C represents the attributor’s (hypothesized use of a) decision-making mechanism, which is the same kind of mechanism employed by the target. Ultimately, Goldman gives up the negative approach and embraces the positive one. His considered view is therefore a ST-TT hybrid. He does not argue that mind-reading involves the use of theory, but, at the same time, does not take this to be a threat to his view. This is because he argues that there is something importantly simulational about mindreading, regardless whether theory plays a role in the process. 12 Goldman (2006), 34. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 4 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? When Goldman makes his case for the mirroring model of simulation-based, lowlevel mind-reading, it is not clear whether he takes the model to be pure ST or a hybrid.13 In either case, arguments that low-level mind-reading entails theoretical inferences present little threat to him. This is because he can easily make the move to a hybrid view. However, he is committed to there being some distinctively simulational features of low-level mind-reading. For Goldman, it seems that this consists in the role of pretend states (the shaded areas) and the attributor’s use of a decision-making mechanism (the hexagon) of the same kind employed by the target. Therefore, we should take these two features to be distinctive of ST (both pure and hybrid). §2. MIRRORING AND SIMULATION-BASED, LOW-LEVEL MIND-READING Goldman argues that mirror neurons support a simulation-based, low-level model of mind-reading.14 He makes his case on the basis of Face Based Emotional Recognition (FaBER). Our ability to recognize the emotions of others based on their facial expressions is an example of low-level mind-reading. This ability is a low-level ability because FaBERs seem to automatically occur below the conscious level. By contrast, our ability to mind-read the extent to which Mr. Crane and Mr. Tees are upset seems to be high-level. FaBER qualifies mind-reading because recognizing that the target’s face expresses a certain emotion involves classifying that emotion as a certain kind of mental state and attributing a mental state to the target. Goldman cites evidence that patients with deficits in the experience of a certain emotion will reliably fail FaBER tests of that same emotion. He argues that whether When Goldman discusses high-level mind-reading, he is clear that he accepts a hybrid view. However, there is nothing in his discussion of low-level mind-reading, which seems most congenial with a pure ST view, to commit him to a hybrid view. 14 See Goldman (2006), Ch. 6 Simulation in Low-Level Mind-Reading: 113-145. For our purposes, ST in this context may refer to either a pure or hybrid view. 13 CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 5 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? TT can explain this is doubtful. He further argues that ST, specifically, a simulation-based, low-level mirroring mind-reading model, can. In short, ST has explanatory power. Empirical evidence shows that deficits in the experience of a certain emotion and selective deficits in FaBER tests of that same emotion reliably occur together. Goldman cites FaBER tests on patients with such deficits. For example, patients with damage to the amygdala, parts of the brain located deep within the medial temporal lobes that play a role in certain emotional responses, reliably fail FaBER tests. Adolphs et al. performed FaBER tests on a patient, S.M., who suffers from a disease, Urbach-Wiethe, which destroyed her amygdala. They found that she reliably failed to recognize facial expressions of fear. Note that the damage to her amygdala resulted in a severe deficit in her experience of fear.15 So it appears that a deficit in the experience of an emotion somehow involves a related deficit – poor performance on FaBER tests of that same emotion. Can TT account for why FaBER deficits of a certain emotion reliably occur with deficits in the experience of that emotion? Goldman argues that patients with such deficits have all of the relevant theoretical knowledge. That is, it does not seem that the patients make an error in how they deploy theory and thereby make the inferences. In the case of S.M., she does not lack visual knowledge of the target’s facial expression, knowledge that such an expression is associated with a certain emotion, or knowledge about that emotion. If TT can offer a plausible account, it must provide some plausible explanation as to why S.M.’s folk-psychological theory about fear went awry on some other grounds. Therefore, Goldman concludes that a TT defender would be hard-pressed to provide an explanation for this very real phenomenon. 15 See Damasio, 66. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 6 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? Can ST provide an account? Goldman argues that it can. His strategy is to examine whether four possible simulation models can explain why FaBER deficits of a certain emotion reliably occur with deficits in the experience of that emotion.16 Since my space here is limited, I will only discuss the one Goldman favors: the mirroring model.17 Mirror neurons ground the mirroring model. Recall that a mirror neuron is a neuron that fires both when we perform some action and when we observe someone else performing some action. So mirror neurons indicate that we have a low-level mirroring mechanism, which may be described as a causal pathway between the target (sender) and the attributor.18 According to Goldman, this pathway has the following two components: (a) a subpath within the sender from his own mental state to a behavioral expression of that state and (b) a subpath within the receiver from an observation of the sender’s behavior to a mental state that matches the sender’s19 Because of this mirroring mechanism, the mental state of the sender (target) is mirrored by a similar one in the receiver (attributor). So the attributor, who receives the mirrored mental state, makes use of this mirroring mechanism for mind-reading. The pathway from (a) to (b) is direct in the sense that it is unmediated by high-level cognition. Indeed, the attributor may be entirely unaware that she adopts mirrored mental states. In the case of FaBER, we can represent this mirroring mechanism diagrammatically as follows: FaBER Mirroring Mechanism The Generate and Test Model: [1] Generate a hypothesis → [2] Produce a facial expression → [3] Test this facial expression against the target → [4] Is there a match? If so, continue [5]. (If not, return to [1]) → [5] Classify one’s current emotional state and attribute this state to the target. See Goldman (2006), 124-125. The Reverse Simulation Model: [1] Visual representation of the target’s facial expression → [2] Activation of facial muscles which imitate target’s facial expression → [3] Experience of emotion → [4] Classification of the current emotion state and attribution to the target. See Goldman (2006), 125-126. The Reverse Simulation with as-if Loop Model: of what it would feel like to make the expression → [3] Experience of emotion → [4] Classification of the current emotion state and attribution to the target. See Goldman (2006), 127. 17 Goldman identifies problems with each of the three models in explaining the FaBER test results; see Goldman (2006), 129-132. 18 See Fadiga et al. and Rizzolatti et al. (1996b). 19 Goldman (2006), 133. 16 CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 7 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? [1] Observing the target’s facial expression. ↓ [2] Having an emotion that mirrors the target’s emotion. This mirroring mechanism, however, does not in itself entail mind-reading. The attributor may potentially draw from her mirrored mental state when mind-reading, but Goldman claims that she must take two additional steps to attribute the mental state to the target. First, she must classify her mirrored mental state token, which she experiences in [2], as an instance of the relevant mental state type. For example, she may feel a fearful emotion, but she may be unable to classify this instance of a fearful emotion as fear. It seems like she must classify her mental state in order to attribute it to the target. Goldman calls this MClassification, “selecting a mental-state category, or classification.”20 Additionally, the attributor must also attribute the mental state to the target. Goldman calls this Imputation, “imputing an instance of that classification to the … target.”21 We can represent the complete mirroring model diagrammatically as follows: Mirroring Model [1] Observing the target’s expression of a mental state. ↓ [2] Having a mental state that matches the target’s mental state. ↓ [3] M-Classification. ↓ [4] Imputation. This model can explain selective failures on the FaBER tests. We can now understand why Goldman claims that the mirroring model explains the face-based recognition deficits. Deficits in the experience of an emotion reliably co-occur with selective deficits in the face-based recognition of the same emotion. This is because of a defect in the mirroring mechanism of that emotion. Step [2] clearly goes awry in the FaBER test patients 20 21 Goldman (2006), 133. Goldman (2006), 133. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 8 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? because of their emotional deficits. Recall that S.M. has a severe deficit in her experience of fear. An attributor cannot complete steps [3] and [4] without first completing step [2]. §3. IS THE MIRRORING MODEL SIMULATIONAL? My goal in this section is not to argue that Goldman’s mirroring model is inaccurate. Rather, it is to argue that Goldman fails to show that it is distinctively simulational. At this point, we may fairly ask Goldman, “Where does simulation occur in the mirroring model?” Unlike Figure C, this model does not involve pretending or a decision-making mechanism (of the same kind employed by the target), which we may recall are the features Goldman seems to take to be distinctive of ST. 22 However, he explicitly concedes that (1) pretend mental states and (2) a decision-making mechanism are not necessary conditions for simulation-based, low-level mind-reading.23 This is because pretending and decision-making are high-level activities which play a role in high-level, but not low-level, mind-reading.24 This is an odd concession, one it is not clear that he can make. In order to argue that mirror neurons support simulational mind-reading, Goldman must at the very least show that there is something distinctively simulational about the mirroring model. Mirror neurons suggest a mirroring mechanism: when the attributor observes the target’s expression of a mental state she will experience a mental state that matches the target’s mental state. I grant this. The evidence from the Parma neurophysiologists is convincing. But does this further suggest anything distinctively simulational about mindreading? A TT defender may well argue that there are distinctively theoretical features concealed in mirroring-based, low-level mind-reading on the grounds that the mirroring Goldman (2006), 34. See Goldman (2006), 131-132. 24 See Goldman (2006), 131-132. 22 23 CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 9 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? model suppresses an intermediary step: [1→2] the inference from observing the target’s expression of a mental state to the attributor’s experience of a mental state. This inference seems to be distinctively theoretical features of the mirroring model, one that draws from some folk-psychological laws. In the absence of distinctively simulational features, this suggests that the mirroring model is more congenial with TT than ST.25 [1] Observation of the target’s expression of a mental state. ↓ [1→2] Inference from the target’s expression of a mental state to the attributor’s mental state. ↓ [2] A mental state that matches the target’s mental state.26 Recall that mirror neurons were first discovered in the F5 ventral premotor cortex of the macaque monkey. The neurons in this area are sensorimotor, i.e., they are responsive to both motor activity and sensory observations thereof. Mirror neurons in humans are also sensorimotor. For example, Fadiga et al. performed a transcranial magnetic stimulation study, which demonstrates that when we observe others performing hand actions, some of our sensorimotor neurons mirror this neuro-motor activity. So the attributor observes motor In a longer draft of this paper, I argue that there are additional inferential steps suppressed in Goldman’s mirroring model. Namely, [2→3] the inference from the attributor’s mental state to a classification of that mental state, [3→4] the inference from the classification of that mental state to the attribution of the target’s mental state. 26 Note that a TT defender has the resources to argue that our perceptual processes themselves are inferentially mediated, regardless of the role they play in mind-reading. In other words, in order to observe the target’s expression of a mental state in the first place, we might think that low-level inferences are necessary. By way of illustration, Fodor tells us the following fairytale about the history of the philosophy of perception. Once upon a time, the Wicked Behaviorist claimed that perceptual processes are reflexes, which are non-inferential and cognitively impenetrable. But the Handsome Cognitivist claimed that perceptual processes are inferential and cognitively penetrable. Inferential mediation means that the perceptual stimuli elicit reflexes indirectly, mediated by cognitive processes. Cognitive penetrability means that the perception is not informationally encapsulated from high-level processes. However, Fodor argues: “Although perception is smart like cognition in that it is typically inferential, it is nevertheless dumb like reflexes in that it is typically encapsulated” (Fodor, 514). So the Wicked Behaviorist was wrong to claim that perceptual processes are non-inferential because of the poverty of stimulus arguments, which maintain that information from our perceptual stimuli radically underdetermines our perceptions (Fodor, 516). Poverty of the stimulus arguments draw from examples such as depth perception and color constancy. So in order for the attributor to even observe the target’s expression of a mental state in step [1], the attributor’s perceptual processes must inferentially mediate information received from the stimuli in her perceptual field. This, of course, is causally prior to her experience of a mental state in step [2]. Therefore, a TT defender may argue that there are hidden inferences in our very perceptual processes, which the mirroring mechanism takes for granted. 25 CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 10 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? activity in step [1]. Then, by way of the mirroring mechanism, she adopts some sort of a mirrored mental state. That is to say, the mirroring mechanism causes some of her neurons to mirror the target’s neuro-motor activity. But does mirroring the target’s neuro-motor activity alone get the attributor to step [2]? On this story, the attributor’s mental state seems entirely devoid of the intentionality of the target’s mental state. In order for the attributor to mirror, for example, the target’s fear of the snake in front of her, it seems that the attributor must in some sense take the snake as the object of her mirrored fear. Whether the mirroring mechanism can account for the intentionality of mental states is controversial. For example, Rizzolatti et al. (2001) make the disclaimer that it is not clear whether the mirroring mechanism is sufficient for mirrored intentionality. They claim: It is important to stress that we are not claiming that … the monkeys understood the intention of the agent (that is, why the observed action was performed), but only that they understood the action meaning (that is, what the agent did).27 Yet, the Parma neurophysiologists seem more optimistic when it comes to humans. Rizzolatti et al. (1996a) suggest that the mirroring mechanism “[enables] implicit action understanding.”28 Further, Iacoboni et al. suggest that mirror neurons are “encoded” with some intentionality. But to what extent this ‘enabling’ or ‘encoding’ can be explained on the neurophysiological level is, at best, questionable. One may indeed argue that motor mirroring is not sufficient for mirrored intentionality.29 By way of example, suppose I reach for a glass of water. We can explain my decision-making procedure by reference to Figure A. That is, I reach for the glass because I have a desire to achieve a goal to quench my thirst and I believe that some course of action, such as reaching for the glass, will achieve that goal. So I plug this belief and desire into my Rizzolatti et al. (2001), 667. Gallese, 173. Also see Rizzolatti et al. (1996a). 29 This same line of argument has been pursued by Jacob. 27 28 CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 11 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? decision-making mechanism, which results in my decision to reach for the glass of water. Further suppose that an attributor observes my behavior. Thus, some of her sensorimotor neurons fire in a way that resembles my own neuro-motor activity. But in what sense does her mental state mirror mine? Merely insofar that some of her sensorimotor neurons mirror mine. So perhaps she mirrors some aspect of my mental state, that is, in an extremely deflationary sense. But this step alone does not seem to help with mind-reading. My attributor does not share my mental state in the sense that I desire to achieve a goal and have a belief that some course of action will achieve that goal. This, it seems, requires an inferential step. She may make use of a folk-psychological law, such as, when someone reaches for a glass of water, he does so because he is thirsty and believes that drinking the glass of water will quench his thirst. In order to have a mirrored mental state, over and above mirrored neuro-motor activity, it seems that she must do some inferential work. So too, we can apply this line of reasoning to low-level mind-reading. The FaBER attributor will have mirrored neuro-motor activity on the basis of the target’s facial expression. But to have a mirrored mental state, over and above mirrored neuro-motor activity, it seems that she must also make some sort of low-level inference(s). 30 Furthermore, Borg has argued that the mirroring mechanism fails to capture intentionality even if the mirror neurons are encoded. On the one hand, the mirroring mechanism is too fine-grained to encode intentionality.31 This is because different mirroring Recent empirical findings buttress this line of reasoning. Busso et. al. set up the following experiment. They use a computer-based model (drawing from an Ekmanian facial action coding system) to present different facial expressions of four different emotions: anger, sadness, happiness and neutrality. They found that attributors correctly classify an emotion 85.1% of the time when presented with a visual representation of the facial expression (Busso et. al., 209). By contrast, attributors correctly classify 89% of the time when presented with both a visual representation and certain acoustic information. They conclude, “Humans use more than one modality to recognize emotions” (Busso et. al., 210). This suggests that visuomotor mirroring is, at best, only part of the story – and surely it is insufficient for mind-reading. It seems that humans, as a matter of fact, draw from multiple modalities to perform FaBERs. 31 See Borg, 14. 30 CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 12 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? patterns can apply to a single intention. For example, suppose I reach for the glass with my left hand instead of my right. The former is a distinct motor activity from the latter. But regardless of which hand I reach with, one may attribute the same intention to me. In either case, my intention is to quench my thirst. On the other hand, the mirroring mechanism is also not fine-grained enough to encode intentionality.32 This is because different intentions can apply to the same mirroring pattern. For example, I may reach for the glass to hand it to the attributor (perhaps because I think that she is also thirsty) or I may simply reach for it to drink myself. We can also apply this line of reasoning to low-level mind-reading. Two different facial expressions convey the same emotion. At the same time, a single facial expression may convey two different emotions. The upshot of this section is that TT has the resources to account for why FaBER deficits of a certain emotion reliably occur with deficits in the experience of that emotion. If we accept that the mirroring model is theory based, Goldman’s argument for simulationbased, low-level mind-reading is no threat to TT. This is because FaBER is not a problem for the mirroring model. TT may accept that some mirroring mechanism is necessary for mind-reading. But this does not point to ST over TT. To sum up, mirror neurons suggest that when the attributor observes the target’s expression of a mental state she will experience a mental state that matches the target’s mental state. We find this in the mirroring mechanism, which is the basis for Goldman’s mirroring model. But we do not find any distinctively simulational features in the model. Note that I did not argue that TT is true or that it offers an accurate model. Rather, I argued for the plausibility of concealed inferences in Goldman’s mirroring model. This suggests 32 See Borg, 15. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 13 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? something theoretical about mind-reading, not simulational. At the very least, Goldman fails to show that there is something distinctively simulational about the mirroring model. §4. CONCLUSION I have argued that the discovery of mirror neurons does not support ST. In the first two sections, I explained Goldman’s distinction between ST and TT and his case for why mirror neurons support simulation-based, low-level mind-reading. In the third, I argued that he fails to show that there are any distinctively simulational features in his mirroring model. Rather, I suggested that the mirroring model has distinctively theoretical features. So where does this leave us? Although there is strong evidence to suggest that some mind-reading draws from a mirroring mechanism, applying the ST-TT distinction to low-level mind-reading does not appear to be helpful. Indeed, hidden inferences in the mirroring model are not devastating to Goldman’s view because he can adopt a ST-TT hybrid one of low-level mind-reading. But should we be content with this move? If Goldman were to adopt such a view, it would seem that the distinction between ST and TT becomes useless to him. Should we be content with calling a hybrid view ST? What do we gain by calling it ST? So too, there does not seem to be something devastating to TT. It is also not clear what ST brings to the table in the first place. What would a hybrid view of low-level mind reading add to a TT account? By contrast, TT brings crucial inferences to the table. So there seems to be little motivation for simulational low-level mind-reading. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 14 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? §5. REFERENCES 1. Adolphs, R., Tranel, D., Damasio, H. and Damasio, A. (1994) “Impaired recognition of emotion in facial expressions following bilateral damage to the human amygdale,” Nature 372: 669–672 2. Baron-Cohen, S. (1995) Mindblindness, an Essay on Autism and Theory of Mind (Cambridge UP). 3. Borg, E. (2007) “If mirror neurons are the answer, what was the question?” Journal of Consciousness Studies 14: 5-19. 4. Busso, C., Deng, Z., Yildirim, S., Bulut, M., Lee, C.M., Kazemzadeh, A., Lee, S., Neumann, U., and Narayanan, S. (2004) “Analysis of emotion recognition using facial expressions, speech and multimodal information,” (Proceedings of the International Conference on Multimodal Interfaces): 205-211. 5. Damasio, A. (1999) The Feeling of What Happens (Harcourt Brace). 6. Di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V., and Rizzolatti, G. (1992) “Understanding motor events: a neurophysiological study,” Experimental Brain Research 91: 176-180. 7. Carruthers, P. (1996) “Simulation and self-knowledge: a defence of theory-theory” in Theories of Theories of Mind (Carruthers, P. & Smith, P.K. eds., Cambridge University Press): 22-38. 8. Gallese, V. (2003) “The Roots of Empathy: The Shared Manifold Hypothesis and the Neural Basis of Intersubjectivity,” Psychopathology 36: 171-180. 9. Gallese, V. and Goldman, A. (1998) “Mirror neurons and the simulation theory of mindreading,” Trends in Cognitive Sciences 2: 493-501. 10. Gazzola, V. and Keysers, C. (2009) “The Observation and Execution of Actions Share Motor and Somatosensory Voxels in all Tested Subjects: Single Subject Analyses of Unsmoothed fMRI Data,” Cerebral Cortex 19: 1239-1255. 11. Goldman, A. (1989) “Interpretation Psychologized,” Mind and Language 4: 161-185. Reprinted in Philosophy of Psychology: Contemporary Readings (Bermudez, J.L. ed., Routledge, 2006): 327-351. 12. Goldman, A. (2006) Simulating Minds: The Philosophy, Psychology, and Neuroscience of Mindreading (OUP). 13. Gopnik, A. and Wellman, H. (1992) “Why the Child’s Theory of Mind Really Is a Theory” Mind and Language 7: 145-171. 14. Hebb, D. O. (1949) The Organization of Behavior: A Neuropsychological Theory (Wiley). 15. Iacoboni, M., Molnar-Szakacs, I., Gallese, V., Buccino, G., Mazziotta, J. and Rizzolatti, G. (2005) “Grasping the intentions of others with one’s own mirror neuron system,” Public Library of Science – Biology 3: e79, 1–7. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 15 DO MIRROR NEURONS SUPPORT SIMULATION-BASED, LOW-LEVEL MIND-READING? 16. Jacob, P. (2008) “What do Mirror Neurons Contribute to Human Social Cognition?” Mind and Language 23: 190-223. 17. Kahneman, D. and Tvesky, A. (1982) “The Simulation Heuristic,” in Judgment under Uncertainty (Kahneman, D., Slovic, P., and Tvesky, A., eds., Cambridge UP): 201-208. 18. Fadiga L, Fogassi L, Pavesi G, and Rizzolatti G. (1995) “Motor facilitation during action observation: A magnetic stimulation study,” Journal of Neurophysiology 73: 2608–2611. 19. Fodor, J. (1985) “Précis of the Modularity of Mind,” Behavioral and Brain Sciences 8: 1-42. Reprinted in Philosophy of Psychology: A Contemporary Introduction (Bermudez, J.L. eds., Routledge, 2006): 513-523. 20. Ramachandran, V. (2000) “Mirror Neurons and imitation learning as the driving force behind ‘the great leap forward’ in human evolution” Edge 69: <http://edge.org/documents/archive/edge69.html>. 21. Rizzolatti, G. and Craighero , L . (2004) “The Mirror-Neuron System,” Annual Review of Neuroscience 27: 169 – 92. 22. Rizzolatti, G., Camarda, R., Fogassi, M., Gentilucci, M., Luppino, G., and Matelli, M. (1988) “Functional organization of inferior area 6 in the macaque monkey: II. Area F5 and the control of distal movements” Experimental Brain Research 71: 491-507. 23. Rizzolatti, G., Fadiga, L., Fogassi, L., and Gallese V. (1996a) “Premotor cortex and the recognition of motor actions,” Cognitive Brain Research 3:131–141. 24. Rizzolatti, G., Fadiga, L., Matelli, M., Bettinardi, V., Paulesu, E., Perani, D., and Fazio, G. (1996b) “Localization of grasp representations in humans by positron emission tomography: I. Observation versus execution.” Experimental Brain Research 111: 246– 252. 25. Rizzolatti, G., Fogassi, L. and Gallese, V. (2001) “Neurophysiological mechanisms underlying the understanding and imitation of action.” Nature Reviews Neuroscience 2: 661 – 670. 26. Stitch, S. and Nichols, S. (1992) “Folk Psychology: Simulation vs. Tacit Theory,” Mind and Language 7: 29-65. CONFERENCE SUBMISSION DRAFT JANUARY 11, 2010 16