Perceiving mental states : Co-presence and constitution

Recently, several philosophers have called attention to the idea that there are occasions on which we can perceive (at least some) mental states of others. In this paper we consider two recent proposals in this direction: the co-presence thesis (Smith, 2010) and the hybrid model (Krueger and Overgaard, 2012). We will examine the aforementioned alternatives and present some objections to both of them. Then, we will propose a way of integrating both accounts which allows us to avoid such objections. Broadly stated, our idea is that by perceiving other people’s behaviors we also perceive their mental states because behaviors co-present some features of the latter, and that this perception of others’ minds is direct and immediate because behavior is a constitutive part of the mental states in question.

of knowing about them: perception.But the mainstream tradition in philosophy of mind has understood this in a different way.In the debate about how we attribute mental states to others, the dominant positions usua ly only concede an indirect role to perception.It is genera ly accepted that perception only gives us knowledge about bodies and behaviors from which, after an a ditional extra-perceptual inferential process, we reach the mental states possessed by the person in question.However, in the last few years, a growing number of philosophers have defended the idea that, as those everyday claims seem to su gest, on some occasions it is possi le to perceive the mental states of others (or at least some of them) in a direct way. 3The basic idea is that, in many of our encounters with other people, we have direct perceptual access to a ects of their mentality similar to the perceptual access to ordinary objects that we enjoy under normal conditions (Krueger and Overgaard, 2012;Krueger, 2012). 4In those cases, the contents of our perceptual states represent the mental states of other people; we litera ly see those mental states in their behavior. 5 This thesis, according to which we can have direct perceptual access to the mental states of others, is frequently ca led Direct Social Pe ce tion (hereafter DSP). 6 Now, according to some philosophers, it is important to distin uish different ways of understanding DSP (Krueger, 2012;McNei l, 2012;Herschbach, 2015;Michael and De Bruin, 2015;Bohl and Gangopa hyay, 2014): as a psychological, epistemological or phenomenological claim.In the first place, DSP can be an empirical, psyc ological claim about what sort of mechanisms ena le us to attribute mental states to others (where it is frequently assumed that these mechanisms wi l be largely tacit or sub-personal). 7In this sense, DSP aims to provide an alternative to traditional theories of mindreading like Theory-Theory (hereafter TT) and Simulation-Theory (hereafter ST).While for the latter we only attribute mental states to others after non-perceptual processes involving theoretical knowledge or simulations, the former claims that it is possi le to attribute mental states by using exclusively perceptual mechanisms.
Secon ly, DSP could be mainly concerned with e istemological issues about how we can come to know the mental states of others, whether we can have perceptual and non-inferential knowledge of them, and what kind of evidence or justification (perceptual, inferential, etc.) we need to have in order to war ant those attributions. 8Under this reading, DSP op oses "inferentialism" (McNei l, 2012;Spaulding, 2015).According to inferentialists, a l our knowledge about the mental states of others is inferential.We can have direct perceptual knowledge about the behaviors and gestures of other people, but we cannot have direct perceptual knowledge of their mental lives.Our knowledge about the mental states of others must be inferentia ly derived from our perception of their observa le behavior and other non-mental features.DSP defenders, on the contrary, ar ue that when perceiving the behaviors of others we can directly perceive their mental states and, at the same time, obtain genuine non-inferential knowledge about them.
Fina ly, DSP could be understood as a phenomenological thesis regarding how other minds are consciously perceived.Those who defend DSP as a phenomenological claim reject a widely accepted assumption: that the minds of others are unobserva le and that, as a consequence, a l we can directly experience are their behaviors.Mentality must be somehow 3 In the analytic tradition, some early advocates of this possibility are Wittgenstein (2009), McDowell (1998), and Dretske (1973).With some differences between them, recent defenders of this approach are Cassam (2007), Krueger and Overgaard (2012), Krueger (2012Krueger ( , 2013aKrueger ( , 2013b)), Smith (2010), McGeer (2009), Gallagher (2008), Zahavi (2011), Stout (2010), and McNeill (2010). 4In the context of the epistemology of perception and philosophy of mind, the expressions "direct perception" and "immediate perception" are considered equivalent, and "direct" and "immediate" are mainly understood in the sense of "non-inferential".See, for instance, McDermid (2001), Malcom (1953), and Snowdon (1992).In particular, Snowdon (1992) elucidates the notion of direct perception as follows: a subject S directly perceives an object O if and only if S stands, in virtue of her perceptual experience, in such a relation to O that, if S could make demonstrative judgments, then it would be possible for S to make the true demonstrative judgment "That is O". 5 In this paper we will focus on the visual perception of mental states.However, we also consider it plausible to think that sometimes we can hear, feel, etc. the mental states of others.It would be possible, for example, that while talking to someone on the phone one could hear her sadness or feel someone's nervousness in her handshake. 6DSP advocates usually focus on intentions and perceptions as the kinds of mental states that can be directly perceived.It is more controversial whether DSP can be also extended to account for the attribution of other types of mental states with more indirect and tenuous links with behavior. 7See Herschbach (2008) and Lavelle (2012) for a defense of a sub-personal reading of the Simulation Theory (ST) and the Theory-Theory (TT).
8 Sometimes philosophers disagree on whether DSP should be interpreted as an answer to the empirical question or to the epistemological one.For example, whereas McNeill (2012) focuses on DSP understood as an epistemological hypothesis, Krueger (2012) seems to be mainly interested in dealing with DSP as an answer to the empirical question (although Krueger, 2013a, also suggests that, once we endorse DSP and deny the unobservability of mental states, the epistemological problem of other minds dissipates).Thus, Krueger holds that the rivals of DSP are, in fact, the Theory-Theory and the Simulation Theory.Smith's position on the subject is more complex.On the one hand, he acknowledges that the epistemological side of the debate is the main concern in his paper.On the other hand, he presents DSP as an alternative to TT and ST, and he argues that, even though they are both usually proposed as descriptive/empirical theories, they can also be stated as answers to the epistemological debate on the inferential/non-inferential status of our knowledge of other minds.His main idea here is that, insofar as the input into the tacit theory/simulation process is the content of a perceptual state, TT and ST defenders will assume that such content does not yet ascribe mental predicates to others (Smith, 2010, footnote 4).Thus, according to them, some additional cognitive process must be added to perception in order to have knowledge of the minds of others.
infer ed from directly experienced pu lic data.DSP advocates point out, instead, that we can have a direct experiential access to the minds of others; i.e., we can perceive their mental states in their gestures and behaviors (Bohl and Gangopa hyay, 2014).
These three ways of understanding DSP somehow reject that perceiving others' mental states involves inferential processes.But, as we sha l see, the notion of inference changes in each of these debates.As mentioned above, when DSP is taken as a psychological claim, it op oses TT and ST.Now, according to the interpretation usua ly given by DSP advocates, TT defenders claim that the inferences in question involve a body of theoretical knowledge about the causal/rational relations between mental states, environmental stimuli and behavior.ST advocates, instead, posit a process of simulation of the likely mental states of the others and a subsequent inference that a lows the attribution of those simulated mental states to them (Herschbach, 2015) 9 .In any case, these inferential processes are genera ly considered to be sub-personal and tacit.They are not open to intro ection, or conscious deliberation, nor do they have any phenomenologica ly distinctive chara er.
In the epistemic debates which revolve around how we can have knowledge of the mental states of others, the so ca led "inferentialists" consider that, in order to show how a subject S could count as knowing that another person C is in mental state M, we need to posit an inference which a lows the derivation of S's knowledge regarding C's mental states from a set of basic propositional perceptual states of S -states about C's behavior, gestures and so on-and some generalizations which connect the latter with the former (McNei l, 2012). 10ence, for the inferentialist, S's knowledge or awareness of C's mental states must always be inferentia ly derived from a more basic awareness of C's non-mental features.On the contrary, DSP defenders contend that we can sometimes have perceptual knowledge of the mental states of others, which is both epistemica ly basic and non-inferential.
Fina ly, when DSP is posited as a phenomenological or experiential claim, it op oses the idea that our experience of the mental states of others is always the experience of forming a belief about them via an inferential process from the conscious perception of their gestures and behavior (Herschbach, 2015).Thus, it rejects the idea that there is an actual inferential process, of which we are somehow aware, mediating between our perceptual states about the gestures and behaviors of others and our beliefs about their mental states.DSP defenders claim, instead, that, at least on some occasions, we directly experience the sadness or the anger in someone's face or in her behavior, instead of acquiring knowledge or beliefs about the sadness secondarily and as a result of perceiving such behavior.
These clarifications are important because, in recent times, many philosophers, neuroscientists, and psychologists have chara erized perception itself as an inferential process (Hershbach, 2015;Michael and De Bruin, 2015).If this is so, a l we can claim, at best, is that there is an indirect and inferentia ly mediated perception of the mental states of others, but never a direct one.11Now, those who think that perception is inferential usua ly endorse a deflated, sub-personal, implicit and non-propositional notion of inference.That is why it seems to us that such a claim is only relevant, at best, for those intere ed in defending DSP as an empirical psychological claim.Now, in this paper, we wi l focus on DSP as a personal-level hypothesis about both the epistemology and the phenomenology of perception, and we think that the claim that perception is inferential loses its relevance once we understand DSP in this sense.The reason is straightforward.In the epistemological debate, the kind of inferential process posed by the epistemic inferentialist, and rejected by DSP advocates, is one composed of propositiona ly structured premises (instead of being a non-propositional process).In turn, advocates of DSP as a phenomenological thesis reject that our access to the mental states of others is always expe ienced as indirect and inferentia ly mediated.Hence, they reject conscious, personal-level inferences as an ineludi le component of our experience of other minds, not tacit sub-personal ones.
Now, when we try to clarify what DSP could consist of, the alternatives proliferate.In what fo lows, we wi l examine and evaluate two proposals: the co-presence thesis (Smith, 2010) and the constitution thesis -or the hybrid model- (Krueger and Overgaard, 2012;Krueger, 2012).More ecifica ly, we would like to examine the merits and limits of these two models, if we think of them as ways of fleshing out the epistemic and phenomenological versions of DSP.We do not aim to ar ue in favor of DSP against its inferentialist rivals.Instead, we would like to explore which would be the best theoretical alternative for those intere ed in adopting this ap roach.In the first section of the article, we present the co-presence thesis and an objection against it.In the second section, we briefly reconstruct the hybrid model of Krueger and Overgaard and, in turn, raise some cha lenges to it.In the third section, we present an alternative proposal that overcomes the criticisms presented against the two previous models and we show how the co-presence thesis and the constitution thesis can be successfu ly integrated.This third proposal intends to incorporate the best elements of the two others, providing us with a more adequate and stronger explanatory model of what perceptual knowledge of the mental states of others could consist in.

The Co-presence Thesis
In his paper "Seeing Other People" (2010), Smith tries to shed light on the thesis that -at least on some occasions-it is possi le to perceive other people's mental states.Smith does not deny that, sometimes, our knowledge of other minds is inferential or indirect.Nevertheless, he believes that this is not always so and, therefore, wants to clarify and defend what he ca ls "The Perceptual View" .To do so, he combines a functionalist account of mental properties with a Husserlian thesis about perception.
Taking into account a we l-known idea from Husserl, Smith points out that there is more to our visual experience than meets the eye.For example, when we visua ly perceive a solid object, like a cup, only a part of it faces us (i.e., only a part of it reflects light onto the retina to form the retinal image).Regar less of this, we expe ience the whole cup as a complete object with a front and a backside, and not merely as a bi-dimensional picture.Generalizing, physical objects are presented to us from a particular per ective, in virtue of certain modes of ap earance that depend on our location in space with re ect to them.Thus, in every perceptual experience of an object there is a core constituted by the a ects of it directly presented to sight.Yet, in each perceptual experience, this presented a ect is accompanied by other features of the object, which are not actua ly seen.These other a ects are co-presented when perceiving the object (they constitute, in Husserl's terms, "the internal horizon" of the perceived object).Thus, when we see a book, only one a ect of it is presented to us: its front.Even though they are not presented, its rear a ect and its innards are -according to Smithco-presented.When we perceive an object, its co-presented a ects are those that we anticipate would be presented to us if, for instance, we changed our location with re ect to the object that we see.In fact, such anticipations have an "if… then…" structure because they refer to the features we would see if we moved or intera ed with the perceived objects in certain ways.These anticipations are genuinely perceptual, rather than being mere beliefs about possi le experiences.Nevertheless, Smith acknowledges that they lack the "intuitional fu lness" of the fu ly presented a ects.
Basica ly, Smith's strategy consists of extending this notion of perception of physical objects to the perception of other people's mental states: Just as the rear aspect of the book is visually present without being visually presented, so another's misery is visually present even though only their frown is visually presented.This view would count as a perceptual account of our access to others' mental states, but would also respect the deep-seated intuition that others' mental states are in some sense hidden from view (Smith, 2010, p. 739).
Briefly, the idea is that while behavior is presented in visual experience, mental states -strictly eaking-are only co-presented.12Now, Smith adopts a functionalist conception of mental properties, according to which properties such as being angry or being in pain are functiona ly individuated in terms of the behaviors that their owners tend to car y out under certain circumstances.Thus, if S has a mental property M, this means that S has a pattern of dispositions to behave in ecific ways when S is in circumstances C or in the presence of certain stimuli I.Moreover, there is a set of conditional sentences of the form "S wi l behave in a certain way B in the presence of stimuli I" which describes such behavioral dispositions adequately.But then, let us sup ose that by perceiving S's behavior we are a le to precisely anticipate that S has those dispositions that individuate M (the mental property expressed by S's behavior).In such a case, we would be accurately grasping the functional role (or the dispositional profile) of M. And this would be -according to Smith-a way of perceiving that S possesses the mental state M. In this way, Smith ar ives at what he ca ls 'principle L' .
(L) For any object O and functional property F, if the perceptual anticipations in one's perception of O 'latch onto' the functional role definitive of F, then one perceives O as being F (Smith, 2010, p. 741).
According to Smith, (L) explains how the functional properties enter into the content of our perceptual experience.The "if… then…" structure, which reflects the perceptual anticipations we form when perceiving something, shows how perception can put us in contact with the functional properties chara eristic of mental states.
Notwithstanding these virtues, the co-presence thesis has been objected by Krueger and Overgaard (2012) and Krueger (2012Krueger ( , 2013a)).According to these authors, one pro lem with Smith's proposal is that the perception of others' mental states cannot be equated to the perception of a tridimensional object such as a tomato or a ta le because, in the latter case, we can move our head, our body, etc., until the occluded parts of the object become visua ly presented to us and not merely co-presented.However, nothing similar hap ens in the case of perceiving the mental states of others.Moving ourselves around, peering more closely or moving the body of the other person, wi l never bring their mental states into direct view.As Smith himself acknowl-edges, these can only be co-presented with the perceived behavior.Thus, they can never acquire the "intuitional fu lness" of the fu ly presented.
Smith's answer to this objection is that our anticipations of co-presented mental states can be confirmed, not by means of subsequent presentations of those mental states as such, but rather by means of new behavioral presentations.What one anticipates, in the case of a co-presented mental state, are subsequent presentations of behaviors that, in turn, would co-present, if they actua ly took place, other mental states related to the first one.Thus, if your frown co-presents ir itation, we can anticipate subsequent behaviors -such as cries-which would co-present related mental states, like anger.As Smith claims: "We can regard one's co-presentations of another's mentality to be fulfi led not, as with the rear a ect of the book, by the co-presented becoming presented, but by the co-presented and presented taking part in a harmonious experience" (Smith, 2010, p. 741).
A pro lem with this answer -Krueger and Overgaard ar ue-is that it leads us to conclude that the mental states of others "rea ly are out of reach of our perceptual experiences.A l we ever rea ly see -have presented to us-is behavior.The mental, though somehow co-presented, is never rea ly gi en as such" (Krueger and Overgaard, 2012, p. 245).Expressing it differently: mental phenomena remain unobserva le in exactly the same way as the backside of the tomato (Krueger, 2012).In consequence, according to these authors, the Husserlian thesis of perceptual co-presence does not adequately explain the possibility of perceiving the mental states of others.

The hybrid model and the constitutive relation: An objection
After criticizing the co-presence thesis, Krueger and Overgaard delineate an alternative way of understanding DSP.Concisely, their proposal rests on a hybrid model, according to which some expressive behaviors are constituti e parts of mental phenomena.
Those who defend DSP typica ly claim that the minds of others are, at least on some occasions, perceptua ly accessi le to us in their expressive behaviors.Nevertheless, Krueger and Overgaard point out that this way of putting things remains ambi uous at a crucial point that needs further clarification.Basica ly, the pro lem is that it is possi le to understand "expression" , as the term is used when it is claimed that behavior expresses mental states, in at least three different ways.
A first alternative is to think that behavior expresses our inner mental states in the sense that the former is caused by the latter.But, as the authors ar ue, if we opt for understanding "expression" in this sense, we do not abandon the old Cartesian idea that mental states are essentia ly private, inner and experientia ly inaccessi le to anyone but their owner.After a l, for such a view, we can only directly perceive the behavioral effects of the mental states of others, but we can never perceive those mental states themselves.It is clear that such notion of "expression" , which equates mental states to unobserva le entities that can only be inferentia ly known, is not the one that defenders of DSP are in need of.
A second alternative is provided by the thesis of co-presence that we examined in the first section.In this case, claiming that certain behavior expresses certain mental states amounts to affirming that when we perceive the expressive behavior of another subject, associated mental phenomena are experientia ly co-presented.Nonetheless, as we have already mentioned, Krueger and Overgaard reject this second option because they believe that, since mental phenomena can only be co-presented but never directly presented to us in our perceptual experience, they end up being "phenomena ly degraded" (Krueger and Overgaard, 2012, p. 244).
Fina ly, the authors hold that there is a more promising option, which consists of understanding "expression" in a constituti e sense.According to this third alternative, the behaviors of others express their mental states because they are prope parts or prope components of them.Mental states have a hybrid structure, composed of both internal processes (neural, psychological and/or phenomenal) and external, spatia ly located, and pu lica ly perceiva le processes (bodily behaviors and gestures).The main idea here is that (at least some) mental states are complex wholes made up of different heterogeneous components with complementary functions.A l these elements must come together and coordinate their re ective functions harmoniously for the subject to have -or to experience-the mental state in question (at least in a fu l-lown sense) (Krueger, 2013a). 13 13 Does this mean that those behaviors which are proper parts of mental state M are also necessary to instantiate M? The problem with such a strong claim is that, on many occasions, we attribute mental states to others even if they do not express them publicly.However, we think there is a better and more modest way of understanding the constitutive link between mental states and behavior (see also Newen et al., 2015, for a similar proposal).According to it, it is possible to have an instance of mental state M even when one does not actually manifest any of the behaviors that are proper components of M and, conversely, it is also possible to perform those behaviors without having M. Nevertheless, there is a natural correlation between both of them, which allows us to say that paradigmatic or typical instances of M do involve those behaviors.The analogy with ordinary physical objects is clear.Imagine a book lacking its cover.That would not stop us from considering that this particular object is an instance of a book.Nevertheless, typical or paradigmatic instances of books have covers, and covers are indeed proper parts of books.Moreover, there seems to be a normative element involved here as well.Hence, when we see a coverless book, we feel justified in judging that something is missing in it.Likewise, imagine that someone who is supposed to be very sad does not show any of the usual manifestations of sadness.One may feel that something is missing here and, even more, one may look for excusing conditions that explain why those behaviors are absent.
Since some of these constitutive elements are pu lic and observa le patterns of behavior, gestures and bodily actions, it is legitimate to conclude that, according to this view, our minds extend beyond our brains and reach out to the surface of our bodies and its behavioral responses (McNei l, 2012).
Krueger and Overgaard sup ort this thesis by offering some empirical evidence in favor of the complementarity between the private and pu lic components of mental states (see also Krueger, 2012Krueger, , 2013aKrueger, , 2013b)).For example, people with Moebius Syndrome are incapa le of facia ly expressing emotions, and this incapability seems to cause a decrease in the intensity of those emotions.Something similar is reported by patients who have received Botox injections, which inhibit the facial expression of emotions.These and other results su gest that neural and physiological processes must coordinate their functions with the relevant expressive-behavioral processes for the subject to be a le to instantiate (fu l-lown) emotional experiences.According to Krueger and Overgaard, something similar is true in the case of intentions and thoughts.14Thus, in a l these cases, when we perceive certain expressive gestures and behaviors in the body of others, we litera ly perceive parts of their mental states. 15his third option, they ar ue, is the best one for those who wish to defend DSP, since it does not treat the mental states of others as Cartesian entities and does not imply either-as the co-presence thesis seems to do-that there is a phenomenological difference between the way in which we perceive behavior and the way in which we perceive mental states.
It is important to highlight the explicit anti-reductionism of the hybrid model defended by Krueger and Overgaard.As the authors themselves point out, if one wants to defend a constitutive link between behavior and mentality, one has two options: (i) To claim that mental states are identical to, and nothing more than, the expressive behaviors that constitute them.This clearly implies a reduction of mental states to expressive behaviors.(ii) To acknowledge that mental states are not equal to those behaviors with which they are, nevertheless, constitutively related.Gestures and behaviors are parts of mental states, but mental states also have other non-behavioral components (phenomenological, neurological, etc.).
Krueger and Overgaard clearly endorse (ii) and, consequently, they can easily avoid the risk of reductionist behaviorism.According to them, some mental states are hybrid in the sense that they are composed of inner and outer processes.The behavioral component of mental states remains distinct from, for example, the related neural a ivity or the phenomenological a ects of an emotional experience.Particular cases of cognition and emotion are instantiated by the complementary coordination of neural, physiological, and behavioral components.
At the same time, positing a constitutive link in this weak sense gives some initial plausibility to the claim that we can directly perceive (at least some of) the mental states of others.In order to see this point, let us focus on cases of ordinary perception of physical objects.In everyday life, we say, for example, that we see an iceberg even when only its tip emerges from the water.In such a case, we see an iceberg by seeing a proper part of it.Likewise, we say that we perceive a house even when, strictly eaking, we only see its front, etc. Analogously, Krueger and Overgaard ar ue, when we see certain expressive behaviors of other people, what we perceive, directly and without intermediaries, are the mental states of which they are proper parts.Now, despite these virtues, the hybrid model faces its own pro lems.According to it, a subject can perceive the mental states of others directly, even though only parts of them -some of their expressive gestures and behaviorsare visua ly accessi le to her.Now, imagine that a subject S is in front of another person C, and S perceives certain gestures and behaviors of C that, in fact, are proper parts of a mental state M. Then two possibilities arise: (a) On some occasions, S wi l perceive C's mental state M in C's gestures and behavior.For example, when S is in front of C and C is frowning, she wi l perceive that C is angry (S wi l see C's anger in C's behavior) and thereby she wi l obtain knowledge (or at least beliefs) about C's anger as such.(b) On other occasions, instead, S wi l perceive C's behavior and gestures and, despite those behaviors and gestures actua ly being part of C's mental state M, S wi l fail to perceive M in those behaviors.Thus, S wi l neither have a conscious experience of M nor acquire perceptua ly based knowledge or beliefs about M as M. To give just one example, when S sees C frowning, she wi l be actua ly facing C's anger, but a l S wi l perceptua ly experience is the frown.S wi l not see that C is angry.
In other words, as seen above, even if some characteristic gestures and behaviors are proper parts of certain mental states, they are not identical to them.But then, even if people can sometimes perceive the latter in the former, they may not always do so.This seems to be what hap ens, for example, when a person unfamiliar with a specific context, or lacking enough experience, is only capa le of seeing the movements of another person, describing them, etc., but is incapa le of identifying which is the mental state that they express.A boy who sees an unknown girl smiling at him and tries to decipher the meaning of that gesture may be exactly in this situation."Why is she smiling?Is she hap y?Is she mocking me?Is she amused?" he may wonder.In such a case, despite clearly seeing that the girl is smiling, the boy fails to perceive the mental state that the smile expresses.
Something similar hap ens in some cases of perception of physical objects.To see this, let us return to the example of the subject S who perceives an iceberg by seeing its tip.In such a case, S only has direct visual access to the tip of the iceberg, but she does perceive the iceberg as an iceberg.Now, it seems that her accomplishment of such a feat involves some knowledge of the mass of ice under water or, at least, that it involves some expectations about its existence.Presuma ly, this is what a lows S to perceive the tip of the iceberg as an iceberg.But her twin, S*, who lacks such knowledge/expectations, wi l be a le to see the tip of the iceberg, but she wi l not perceive it as (a part of ) an iceberg.S* wi l be deprived, then, of the possibility of perceiving the whole object.Now, it is not clear how Krueger and Overgaard's proposal could a low us to differentiate the kind of situation described in (a) where S perceives C's mental state in her overt behavior from (b) where S perceives C's behavior as mere behavior.In both situations we have a subject who perceives a proper part -some behavior-of a larger totality: the mental state.But why does it hap en that sometimes the perceiver can have (some) epistemic perceptual access to the totality while, on other occasions, a l she perceives is the proper part?Besides, how can we explain the difference between these two cases?Since Krueger and Overgaard are intere ed in elaborating DSP as a claim about the possibility of having perceptual experience and perceptual knowledge of the mental states of others, they need to elaborate an answer to these questions, but their account does not provide us with the necessary elements to do so.
In summary, Krueger and Overgaard owe us an explanation of how a subject who has visual access to some expressive behavior can have also a perceptual access to the mental state of which such behavior is a proper part.And, at this point, saying that some expressive behaviors are constitutive parts of mental states does not help at a l, because a constitutive relation -as they chara erize it-is just an ontological link that relates behaviors to mental states independently of any knowledge we may have of the aforementioned relata.

Co-presence, constitution and partial knowledge of other minds
In this section, we wi l try to show that Smith's proposal is a valua le tool if we want an account of DSP which a lows us to distin uish perception of another person's mental states from perception of her cur ent behavior.However, we sti l need to give an ap ropriate answer to Krueger and Overgaard's objection against the co-presence claim presented above, in the section devoted to presenting Smith's proposal.We wi l therefore ar ue, in what follows, that it is also possi le to provide such an answer if one is wi ling to combine the co-presence thesis with the hybrid model defended by its critics.Thus, by articulating the co-presence thesis and the claim that there is a constitutive link between mental states and behaviors, we wi l both give a response to Krueger and Overgaard's objection against the former and provide an adequate explanation of what is required in order to perceive the mental states of others as suc in their behaviors.
First, let us go back to Krueger and Overgaard's objection.The defender of co-presence ar ues that perceiving the mental states of others in their behavior is analogous to perceiving a tomato by seeing its front.In both cases, there is something that is visua ly presented to us (the behavior/the front side of the fruit) and hi den a ects that are merely anticipated or co-presented (the mental phenomena expressed in their behavior/the backside and the innards of the fruit).The pro lem is that the analogy fails at a crucial point.When we perceive three-dimensional opaque objects, we can move our bodies in different ways until what was merely co-presented becomes presented.The mentality of others, in contrast, can never be brought into a direct view in the same way.Of course, we can, as Smith claims, "confirm" that we have adequately perceived another person's mental state in her harmonious subsequent behavior.However, a central difference remains: while in the case of perceiving a tomato we can eventua ly come to see its backside directly, when we are facing an angry person, a l we ever rea ly see are simply different behavioral manife ations of her anger in different circumstances.
Thus, as Krueger and Overgaard ar ue, since a l we ever directly see -a l that is ever presented to us-are others' behavioral manife ations, we can doubt "whether CP [the co-presence thesis] makes any advance beyond more traditional accounts, according to which the mental states of others are 'unobserva le' and thus must be infer ed" (Krueger and Overgaard, 2012, p. 245).
How should we a dress this objection?A plausi le reaction would be to reject Krueger and Overgaard's objection on the basis that it is unjustly o livious to the fact that, according to Smith's account, co-presented a ects do constitute part of the content of perceptual experience.Thus, in what sense would it be fair to go on claiming that those mental states remain unobserva le?Now, although this initial reply has some bite, we think that Krueger and Overgaard may sti l have a valid point.To see why this is so, imagine that we asked defenders of the co-presence model whether they think that those behaviors that are expressive of mental states are also proper parts of them or not.If these philosophers deny that expressive behaviors are proper parts of mental states, then they are committed to treating the latter and the former as "different existents".Now, according to Smith's model, our visual experience of others seems to be exhaustively composed of: (i) their presented actual behaviors; and (ii) the co-presented potential behaviors that are not being actua ly manife ed but that, we anticipate, could be manife ed under ap ropriate circumstances.Then, a l we can experience perceptua ly -a l that is ever presented or co-presented to us-are (actual or potential) gestures and behaviors.But if this is so, under the assumption that mental states might be something different from gestures and behaviors, it seems fair to conclude that Krueger and Overgaard are right to complain that, strictly eaking, we never have direct perceptual access to them.
Defenders of co-presence could avoid this perplexing conclusion, of course, by claiming that gestures and behaviors are proper parts of mental states.But this implies combining the co-presence thesis with the constitution thesis proposed by Krueger and Overgaard.Although Smith says nothing about it, it is worth noting that, in principle, there is no incompatibility between these two theses.Nothing in the co-presence thesis prevents us from thinking that the presented aspects of mental states of other people -their behavioral manifestations-could be constitutive parts of those mental states.In fact, claiming that gestures and behaviors can be proper parts of mental states a lows us to keep a tighter analogy with the way in which, according to Husserlian philosophers, we perceive physical things.After a l, when we perceive a tomato, the presented aspect -its front-and its co-presented aspects -its back and its innards-are a l proper parts of the fruit.Likewise, one could consider manifest behaviors and gestures to be the presented proper parts of a larger whole: the mental state that they express.
What we would like to su gest, then, is that the weakness of the co-presence thesis resides not in how it explains what it means to have a perceptual experience, but in the ontological assumption that mental states are something entirely different from the gestures and bodily behaviors that express them.Thus, we believe that by integrating the hybrid model with the co-presence thesis we can obtain a better version of the latter, one that is more consistent with DSP than Smith's original proposal.But let us develop this idea a little more.
On the one hand, according to the hybrid model, gestures and behaviors express mental states because they are a constitutive part of them.On the other hand, the co-presence thesis claims that direct perceptual presentation of others' be-havior is accompanied by other a ects of their mental states which are anticipated by means of "if… then…" structures.By combining both theses, it can be claimed that the presented (and the co-presented) a ects of others' mental states are constitutively linked to the whole mental state.And it is precisely in virtue of that constitutive link that when perceiving an a ect of the mental life of another person (the behavioral a ect), we perceive a part of he mental state.
Even more, according to Smith's version of the co-presence thesis, whilst the gestures and behaviors of others can be visua ly presented to us, their mental states -their anger, their surprise, etc. -are merely co-presented.But now, in virtue of the su ge ed combination of the co-presence thesis with the constitutive thesis, we should substantia ly revise such a claim.When we are looking at someone's typical anger behavior, her anger is not merely co-presented to us.Rather, a part of her anger -the agent's behavior-is already presented to us, while another part remains co-presented.Thus, if we accept this combination of the co-presence thesis and the hybrid model, Krueger and Overgaard's original ar ument against Smith loses its force, as they can no longer claim that the mental state of another person is never presented to us in perception.The reason is straightforward: since expressive behaviors are proper parts of the mental states in question, when those behaviors are visua ly presented to us, the cor esponding mental state is at least partia ly presented to us as we l.Moreover, as has been repeate ly pointed out, what is co-presented to us when seeing an angry person's cur ent gestures and behavior, what we anticipate, are other potential gestures and behaviors which, if they actua ly took place, would be proper parts of her mental state of anger.But, of course, if these potential behaviors actua ly occur ed, they would become perceptua ly presented to us.Therefore, if this hap ened, the mental state constitutively linked to them would not be merely co-presented, but also (partia ly) presented in them.
To be clearer, our su gestion consists of combining a thesis about what perceiving the mental states of others consists in (the co-presence thesis) and an ontological thesis about the nature of those mental states of an agent that can be directly perceived by others (the hybrid model).Of course, the co-presence thesis, such as Smith elaborates it, presup oses an ontology of mental states that is quite different from the one that is held by Krueger and Overgaard.Smith seems to be committed to two different claims about the nature of mental states.On the one hand, he explicitly endorses a functionalist conception of mental states.On the other hand, he seems to accept the idea that mental states are internal and in some sense hi den from view (see Smith, 2010, p. 739).Now, it should be highlighted that even if many philosophers find it natural to combine their functionalist leanings with the claim that mental states are private, unobserva le, and entirely constituted by internal properties, the former claim is, strictly eaking, independent of the latter.And it is only the claim about minds being internal and unobserva le that should be abandoned in order to adopt the kind of hybrid view that we are presenting here. 16 Our proposal concerning the compatibility of these two per ectives relies on the possibility of combining Smith' s theory of how we can perceive the mental states of others and Krueger and Overgaard' s thesis about the ontology of those perceived mental states.The co-presence thesis provides a promising model to understand what perceiving others' mental states can be, but fails to complement this with an ap ropriate conception of what mental states are.In contrast, the hybrid model provides a suita le conception of mental states (at least if one is intere ed in holding DSP), but fails to complement this with an adequate theory of perception.The solution to these shortcomings that we are offering here consists of retaining what we consider to be the best part of each model and dismissing the elements that, although present in them, constitute an impediment to elaborate a more coherent version of DSP.Now, it may be ar ued that, even if the su ge ed integration of the co-presence and the constitutive theses were accomplished, mental states would sti l end up being "phenomena ly degraded" , since there may be some other a ects of another's anger -like its subjective dimension or the way it feels to be angry for that person-which would never become visua ly presented to us.However, the present proposal is prepared to explicitly acknowledge this point because the idea is not that we can perceive every a ect of the mental states of others.Something analogous hap ens, anyway, with the perception of ordinary objects.Just as we cannot perceive every a ect of a fruit that is ripe (we cannot perceive its atomic particles, for example), 17 we are also not a le to perceive the subjective dimension of others' mental states.Notwithstanding, it is sti l true that, if DSP is cor ect, we can perceive the mental states of others.Just as we can see that a fruit is ripe -although we cannot see every a ect of it-we can see that another person is angry -although we cannot see the subjective, or the neurophysiologic, dimension of her anger.
One last thing to note is that we are drawing the analogy between perceiving mental states and perceiving ordinary objects in a slightly different -and, to our mind, more accurate-way than the one Smith offers.To see this, let us return to the case of seeing a tomato.When we see a tomato, its front is presented to us while its back remains co-presented.Smith claims, analogously, that when we see that someone is angry, her behavior is the presented a ect, while the mental property -the anger itself-remains co-presented.But there are two a ditional points that can be made about the experience of seeing a tomato that Smith's analogy does not sufficiently take into account.First, what we experience is the tomato as a whole.Secon ly, there are a ects or properties of the tomato (like its molecular properties) that are not perceived at a l.Our proposal a lows us to include these further a ects by saying that when we see someone's anger, we are presented with a part of it -certain behaviors-while other parts of it -a range of potential behaviors-remain co-presented.But, at the same time, what we are experiencing is the larger whole -the mental state of anger-of which the co-presented a ects are part.And of course, there are other a ects of this larger whole, such as its neurological or subjective properties, which are not part of our perceptual experience at a l.
Let us turn now to the objection to Krueger and Overgaard's hybrid model that we presented above.According to it, to posit a constitutive relation between behaviors and mental states is not sufficient to explain how we are a le to perceive those mental states as suc merely by perceiving their behavioral manife ations.Now, this pro lem can also be solved by combining the hybrid model with the co-presence thesis.Consider the case of a person who frowns.What the defender of co-presence could now te l us is that in order to perceive that gesture as an expres ion of pain, it is necessary to take into account other a ects of that mental state, which, though not presented, are co-presented.We see the frown as a gesture 16 Stricto sensu, functionalists are committed to the idea that what makes something a particular mental state is the way it functions, or the role it plays, in the larger system of which it is part (Levin, 2013;Wheeler, 2010).A traditional way of understanding the functionalist proposal may add to this that these mental states should be understood, metaphysically speaking, as internal entities that -at least in the case of human beings-are identical to or supervene on their owners' brains.In consonance, this kind of view will take the links between these internal mental states and external behavior to be merely causal.Now, this view can be contested.Thus, for example, the so called "liberal" or "extended" functionalist claim that mental states may extend beyond the body, coming to include different items in the creature's environment (Wheeler, 2010;Clark, 2008).The core functionalist idea that mental states have a constitutive, distinctive cognitive role or function is still part of this picture.But, now, the mental states that have this distinctive functional role can be realized in complex and heterogeneous material substrata -that may include internal states of the nervous systems, bodily states, processes and responses, and different entities in the external environment.Thus, mental states are better characterized as complex wholes constituted by these different components and their interactions.Analogously, it seems to us that it is possible to adopt a similar view, according to which at least some mental states extend beyond the brain, including the body and its behaviors (even though they may not include other items in the creature's environment).Advocates of such an embodied view should not understand mental states as purely internal (neural) items that interact in a merely causal way with gestures and behaviors.Rather, they might be thought as complex cognitive states composed of both certain internal states and some gestures and behaviors.And, even though, admittedly, theses different parts will interact causally in many ways -as many components of complex systems do-, these causal relations should be thought as relations amongst parts of these mental states.Consequently, both the causal relations between internal states and behavior and the relata in question will be constitutive components of these mental states.It seems to us that this kind of functionalism is perfectly compatible with the hybrid view that we are trying to defend here. 17At least we do not see them in the sense that is relevant to the present debate: that of being able to perceptually experience them as atomic particles.
of pain because we can anticipate further behavior associated with it.That is what distin uishes perceiving the mental state of another person from merely perceiving her cur ent behavior.In the first case, the perceiver locates the perceived behavior within a broader rational pattern of potential behaviors, and this is what a lows her to perceive those behaviors as part of a larger whole: the mental state of the agent. 18If, on the contrary, a subject is capa le of perceiving another agent's behavior but is incapa le of making the ap ropriate anticipations of relevant future behavioral patterns of the agent, she wi l not be a le to perceive that behavior as an expression (or as a part of) the agent's mental state.Thus, if we are wi ling to combine the co-presence thesis with the hybrid model, we wi l be in a position to solve the difficulties that, as we ar ued in the previous section, undermine the latter.
To sum up, we believe that the best alternative for the defender of DSP is to combine an ontological thesis, which posits a constitutive link between actual behavior and other a ects of the mental phenomenon, and an epistemic thesis -that of co-presence-which plausi ly explains how perception directly provides knowledge of the mental states of others as such.

Final remarks
In this article we have tried to answer the question: which would be the best theoretical alternative for those intere ed in defending the thesis according to which we can directly perceive some mental states of other people?Taking into account two recent proposals -the co-presence and the constitution theses-, we have shown how their re ective weaknesses can be overcome by combining both of them in a complementary way.The cor ect articulation of those proposals gives us, we think, a better way of understanding what direct perception of other minds could consist in.