Are Propositional Attitudes Mental States?

I present an argument that propositional attitudes are not mental states. In a nutshell, the argument is that if propositional attitudes are mental states, then only minded beings could have them; but there are reasons to think that some non-minded beings could bear propositional attitudes. To illustrate this, I appeal to cases of genuine group intentionality. I argue that these are cases in which some group entities bear propositional attitudes, but they are not subjects of mental states. Although propositional attitudes are not mental states, I propose that they are typically co-instantiated with mental states. In an attempt to explain this co-instantiation, I suggest that propositional attitudes of minded beings are typically realized by mental states.


Introduction
If group entities such as companies and organizations could have beliefs and desires, what would that tell us about the nature of beliefs and desires? I argue that it would show that beliefs, desires and more generally propositional attitudes are not mental states. The emphasis in this negative claim is not on the status of propositional attitudes as states, 1 but on their status as mental. "Anti-psychologism" is the name I chose for this view given its implication that propositional attitudes are not part of psychological reality. The argument for anti-psychologism, in a nutshell, is that if propositional attitudes are mental states, then bearing propositional attitudes requires being minded; but there are possible (and perhaps actual) bearers of propositional attitudes that are not minded beings, so propositional attitudes are not mental states.

3
More precisely, I shall argue that: (A1) Some group entities are (or can be) bearers of propositional attitudes. (A2) A property/state is mental only if it can only be had by minded beings. (A3) A being is minded only if it is of such a kind that there is something it is like to be it. (A4) No group entity is (or can be) of such a kind that there is something it is like to be it.
From these four premises, anti-psychologism follows. Sects. 2-4 will expand on this argument, defending its individual premises. If anti-psychologism is true, then propositional attitudes are not part of psychological reality, but this doesn't mean that they have no role in our mental lives. Sect. 5 will consider objections, offer replies, and finally elaborate on how propositional attitudes of minded creatures are intertwined with the mental states of such creatures, arguing that propositional attitudes are paradigmatically co-instantiated with mental states; in minded beings, occurrent propositional attitudes are typically realized by mental states.

Genuine Group Intentionality
The aim of this section is to make a prima facie case for the first premise of my argument for anti-psychologism: (A1) Some group entities are (or can be) bearers of propositional attitudes. This is an endorsement of what I shall call genuine group intentionality (GGI). Here, "genuine" emphasizes two things. First, when there are instances that verify GGI, it is genuinely the groups-not merely the individual members of such groups-that have beliefs or desires. The belief/desire attributions to such groups in this sense are neither metaphorical nor merely derivative over the belief/ desire attributions to some or all individuals of such groups. Second, in cases of GGI, the relevant states are genuinely, in the fullest sense of the term, propositional attitudes. At least some group-beliefs and group-desires are beliefs and desires simpliciter, and they can be type-identical with propositional attitudes of more typical subjects of such states. Although the argument for anti-psychologism that I will present requires only the possibility of GGI, I will try to motivate the view that there may be actual instances of GGI.
Before I proceed any further, let me clarify that defending GGI is not my ultimate goal here. As far as the argument of this paper is concerned, GGI serves a goal in understanding the nature of propositional attitudes. Thus, a full-fledged defense of GGI is beyond the scope of this paper. For recent comprehensive defenses, see Bird (2010), Theiner & O'Connor (2010), List & Pettit (2011), Huebner (2014) and Tollefsen (2015.
A case in favor of GGI can be given as follows. We sometimes attribute propositional attitudes to groups, and in some cases, such attributions are instrumental in understanding and predicting these groups' behaviors. This doesn't constitute a conclusive argument for GGI, but it is a step in the right direction. After all, when we attribute propositional attitudes to other individuals, the fact that such attributions 1 3 Are Propositional Attitudes Mental States? are instrumental in understanding and predicting their behaviors is often a good indicator that they do have these states. 2 Of course, this instrumentalist attitude alone doesn't warrant realism regarding propositional attitudes, either in groups or individuals. But, if we can supplement such instrumentalism with a theory of propositional attitudes which demystifies such states, then we can move towards a more realist view about such states. In the case of the individual, we have every reason to think that this is not an impossible task. There are several theories that demystify, more or less successfully, what it is for an individual to have a belief, for example. 3 If such demystification can be done in the case of groups too, we can be warranted in GGI. Importantly, if our demystifying theory of what it is for some entity to have some propositional attitude can be extended to cover groups without additional ad hoc commitments, we can furthermore make a case for the claim that at least some group-beliefs and group-desires are beliefs and desires simpliciter. I happen to think that functionalism about propositional attitudes does exactly this, and henceforth, I will assume that this theory is true. 4 Although there are different versions of functionalism, 5 there is one core claim in virtually all functionalist theories: the relevant target properties are functional properties in the sense that they are individuated in terms of causal roles, and to have a functional property is to have some property that occupies the relevant causal role. Standardly, functionalist theories are presented alongside a "multiple realizability" claim with the suggestion that the very same functional property F is multiply realizable: in different systems, F can be instantiated in virtue of different properties that occupy the causal role that individuates F. Importantly, this multiple realizability claim is not an additional ad hoc commitment: the possibility of multiple realization is typically seen as an argument for functionalism in the first place (Putnam, 1975). Functionalism about propositional attitudes is then minimally the claim that to have a propositional attitude is to have some property that occupies the causal role that individuates that propositional attitude.
Thanks to multiple realizability, it is not very difficult to get a possible case of GGI from functionalism. Focusing on beliefs for the moment, consider the belief that it will rain this afternoon. Plausibly, the causal profile of this belief will include being caused by receiving relevant information about the likelihood of rain, and causing the production of reports that carry the information that it will rain this afternoon. If functionalism is correct, to the extent that I can form this belief (because I can receive the relevant information about the likelihood of rain, and I can report that it will rain), a group entity, for example the Weather Forecast Agency (WFA), can have this belief too. 6 In my case as an individual being, the properties that play the relevant causal role are neurophysiological properties. In the case of WFA, the relevant property will be a structural property that is instantiated by WFA as a group entity. 7 Henceforth, I will take (A1) to be true. I am hopeful that even those who reject it will be happy to read on to see what GGI entails with respect to the nature of propositional attitudes.

Mental States and Being Minded
In this section, I will present the following two premises of my argument for antipsychologism: (A2) A property/state is mental only if it can only be had by minded beings; (A3) A being is minded only if it is of such a kind that there is something it is like to be it. 8 There is a very straightforward case for (A2). Just as a physical property can only be had by physical beings, and a biological property can only be had by biological beings, a mental property can also be had only by mental, or minded, beings. A qualm about (A2) might be with the notion of a "minded being". But, at least as far as (A2) is concerned, the talk of minded beings shouldn't commit us to any particular ontology of minds. For example, we could take being minded to be nothing over and above having mental properties, which is compatible with (A2).
Let me elaborate on (A3). The sense of "something it is like" that is used here should be familiar: it characterizes subjective/qualitative/phenomenal aspects of conscious experiences. Very few philosophers deny that at least some mental states are such that there is something it is like to be in them. Standard examples of such phenomenally conscious states include sensory perceptions, bodily sensations, pains, pleasures and emotional feelings. In all of these examples, when the relevant 1 3 Are Propositional Attitudes Mental States?
properties are instantiated, there is something it is like for a subject to be in such states.
If (A3) is true, the notion of mindedness is crucially linked to the capacity of having subjective experiences: only those beings that are capable of phenomenal consciousness are minded beings. The question that (A3) can be seen as an answer to is the question of where to draw the boundary between minded and non-minded beings. While one view is that this boundary is between beings that have intentional states and beings that don't have intentional states (see Brentano, 1973 for the view that the mark of the mental is intentionality), the view that I find more plausible is the one that says that the mark of the mental is phenomenal consciousness. 9 On this proposal, minded beings are those that are capable of having conscious experiences. Arguably, the capacity for phenomenal consciousness is at least sufficient for being minded, which then makes it a candidate for marking the boundary for mindedness. Once phenomenal consciousness is a candidate, it is a legitimate working hypothesis that the capacity for phenomenal consciousness is the mark of the mental; it is at least as legitimate as any other (non-disjunctive) candidate. What makes phenomenal consciousness a better candidate compared to intentionality in my view is that what motivates the mind-body problem, what makes the problem interesting, and perhaps intractable, is the peculiarity of phenomenal consciousness (Chalmers, 1996;Strawson, 1994). 10 One might object that (A3) begs the question for anti-psychologism because it ultimately comes down to the claim that only phenomenally conscious states are mental states. However, this is not true; (A3) doesn't say that only phenomenally conscious states are mental states. Rather, (A2) and (A3) together entail that only those states whose instantiations require the capacity for phenomenal consciousness are mental states. That only phenomenally conscious states and states that require phenomenal consciousness are mental states doesn't entail anti-psychologism; it is logically possible-and according to some, actually the case-that propositional attitudes are either phenomenally conscious states or they require phenomenal consciousness. 11 So, neither (A3) nor the conjunction of (A2) and (A3) can be said to beg the question in favor of anti-psychologism. Rather, what we have is 1 3 that these two and further two premises ([A1] and [A4]), together, deductively entail anti-psychologism.
I have argued that a property is mental only if it can only be had by entities which are of the kind that there is something it is like to be them. If an entity that doesn't have the capacity for phenomenal consciousness could instantiate a property P, then P is not a mental property. This doesn't mean that only phenomenally conscious states are mental states. Rather, it means that only those states having of which requires the capacity of having phenomenally conscious states are mental states. Such states (trivially) include phenomenally conscious states, and it is an open question as to whether there are other states which are not themselves phenomenally conscious, but for some reason, require phenomenally conscious states. If there are such states, then the argument of this section doesn't show that they are not mental states.

Against Group Phenomenal Consciousness
I will now argue for the following premise of the argument for anti-psychologism: (A4) No group entity is (or can be) of such a kind that there is something it is like to be it. As discussed in the previous section, the "something it is like" phrase here picks out phenomenal consciousness. So, this premise amounts to the claim that there is no such thing as group phenomenal consciousness; groups cannot be subjects of conscious experiences.
Before giving an argument for (A4), I should clarify one thing. If we understand individual human persons as group entities, surely (A4) is false. But what might be the reason to think that an individual human person is a group entity? One might argue that individual human persons are group entities because they are constituted by groups of entities (e.g., groups of cells). But this is a very liberal understanding of "group entity", and it implies that any composite object is a group entity, which trivializes the question of group intentionality. Thus, we should not understand individual human persons as group entities.
I will motivate (A4) by arguing as follows: (i) that there is no such thing as group phenomenal consciousness is the default position; and (ii) there are, to the best of my knowledge, no persuasive arguments against this default position. Clearly, this argument doesn't even attempt to be deductively sound, but I believe (i) and (ii) strongly support (A4).
In favor of (i): I think it is clear that this really is the default position (even if turned out false). 12 Many contributors to the debate on GGI-proponents and opponents alike-agree that it is the default position. On the one hand, many proponents of GGI explicitly reject the idea that groups have subjective experiences (Theiner & O'Connor, 2010;Gilbert & Pilchman, 2014, pp. 191-192;Tollefsen, 2015, p. 53;List, 2018). On the other hand, opponents of GGI reject this idea for similar reasons they reject GGI (Rupert, 2004, p. 404;Wilson, 2004, p. 294;Baddorf, 2017). Moreover, proponents of group phenomenal consciousness (to be discussed shortly)

3
Are Propositional Attitudes Mental States? acknowledge that they are arguing against a default view. For these reasons, it is uncontroversial that (A4) is the default position, even if it may turn out to be false.
For (ii), namely the claim that the arguments against this default position are not successful, I will consider three such arguments.

The Argument from Crowd Phenomena
The first of these three arguments goes as follows. Some groups exhibit some emotional mental states which are over and above the emotional states of the members of such groups. This line of thinking can be found discussions of some social phenomena, for example in cases where "crowds" are attributed emotional mental states (e.g., Le Bon, 1895). Alleged examples of such phenomena include the anger of a mob and the joy of a cheering crowd. Now, if emotional states are phenomenally conscious mental states, these will be examples whereby some groups are bearers of phenomenally conscious states, therefore, there can be cases of group phenomenal consciousness.
There are two things to say in response to this argument from crowd phenomena. First, it is dubious that the relevant emotional states in these examples are phenomenally conscious states. Emotions can be correctly associated with phenomenal qualities, but we can reserve the term "emotional feeling" (as used as an example for a phenomenally conscious state in Sect. 2 above) for the qualitative aspects that are accompanied by emotions. It is plausible to think that in crowd emotion examples, the relevant states are more akin to propositional attitudes, which can be given a functionalist explanation as per Sect. 2. Then it would be acceptable to attribute such states to groups along the lines of GGI without attributing them phenomenal consciousness. Second, even if we want to stick to the phenomenal aspects of emotions and talk of emotional feelings that are exhibited in such groups, it is not so clear if it makes sense to take groups as subjects of such qualitative states. Such phenomena can be explained perhaps in terms of "feelings of unity" in individual members of groups, where individuals have certain phenomenal states that they wouldn't have if they weren't members of relevant groups (Gilbert, 1989, p. 223; see also Gilbert & Pilchman, 2014). For these reasons, I think the argument from crowd phenomena doesn't work.

The Argument from Combination
The second argument in favor of group phenomenal consciousness is motivated by panpsychist considerations, where panpsychism is the view that phenomenal consciousness is ubiquitous in nature. 13 This argument is from Luke Roelofs (2019), 13 There are many varieties of panpsychism. See Chalmers (2015) for a helpful categorization of panpsychist views. While my discussion in the subsequent three paragraphs will surely not do justice to such variety (see REDACTED for a more detailed treatment), I believe it will suffice for the sake of discussing the relevant argument. For what it is worth, the variety of panpsychism that I discuss here is sometimes called "constitutive panpsychism", and it holds that consciousness of macro-level entities like human beings are constituted by the consciousness of micro-level entities that make up such macro-level entities. who doesn't presuppose panpsychism as a premise, but has a panpsychist-friendly strategy. Roelofs's argument is that the reason we tend to resist the idea of group phenomenal consciousness is the same reason for which we find panpsychism untenable: namely the combination problem (Seager, 1995). Roelofs argues that the combination problem is not really a problem, and by so doing, aims to show that group phenomenal consciousness faces no special problem.
What is the combination problem? If panpsychism is true, then fundamental bits of matter that make up phenomenally conscious creatures like typical human beings are also bearers of phenomenal consciousness. In fact, this is how panpsychism is meant to be a solution to "the hard problem" of consciousness, where the hard problem is that of explaining how bits of matter, say, in one's brain, which are purely physical and not bearers of phenomenal consciousness, make up of phenomenally conscious creatures. The panpsychist solution to the hard problem rejects the assumption that such bits of matter are not phenomenally conscious, and supposedly removes the mystery. The combination problem is the problem of explaining how such relatively simple experiencing beings can combine to generate further, less simple, experiencing beings. Roelofs argues that there is no special problem of combination of experiences. If the problem is motivated by the assumption that no experiencing being has, as a proper part, an experiencing being, that assumption is false-or so argues Roelofs. For example, I am a phenomenally conscious being, but I do have a proper part which is also a phenomenally conscious being, namely my brain (or some proper part of my body which includes my central nervous system and perhaps some peripheral parts). Once we have this kind of "trivial combination", it is easy to see that there is no metaphysical problem regarding combination. Simply imagine the possibility of two (or more) conscious brains that are proper parts of one single experiencing being. So conscious beings could combine to make up conscious beings. 14 I am not sure to what extent I am happy to grant the example from trivial combination, but its status bears little relevance to the case against group phenomenal consciousness. I don't think I need to take a stance on the combination problem for panpsychism. I don't think panpsychism is true, but my skepticism is not guided by the combination problem. Combination may or may not be a problem. I simply reject Roelofs's claim that our resistance to the idea of group phenomenal consciousness is due to the combination problem.
Consider Block's (1980) "nation-brain" thought experiment as a challenge to functionalism about phenomenal consciousness. In this thought experiment, we are assuming that individual citizens of a nation are all recruited to do tasks which are functionally isomorphic to the tasks that neurons or groups of neurons carry out in the brain of a phenomenally conscious creature. The intended conclusion is that it is counterintuitive that such a nation would be subject to conscious experiences.
If Roleofs is right that our resistance to group phenomenal consciousness is due to the combination problem, then the following slightly altered version of this thought experiment shouldn't lead to the same intuition as the original one. Suppose, all citizens of the nation in question are zombies, lacking phenomenal consciousness. In this version of the thought experiment, it is not more plausible (that is, relative to the original version) that such a nation has phenomenal consciousness. That is, the intuition that a group entity like the nation-brain wouldn't instantiate phenomenally conscious properties is not motivated by the intuition that phenomenal consciousness doesn't combine. Then, resistance to group phenomenal consciousness is not due to a worry about members of such groups being phenomenally conscious. Therefore, the combination problem, in the case of group phenomenal consciousness, is a red herring, and the argument from combination fails.

The Argument from Information Processing
The third argument for group phenomenal consciousness is from Eric Schwitzgebel (2015), and it goes as follows. 15 The level of information processing we find in the nervous system of a phenomenally conscious creature (like a typical human person) is similar to the level of information processing that takes place in the internal affairs of some groups, for example the United States as a nation. If phenomenal consciousness is going to be explained in terms of information processing in the brain in the case of a typical human subject, nothing but neuro-chauvinism-i.e., the view that only neural substrates can be seats of conscious experiences-rules out the possibility that the Unites States is a phenomenally conscious entity too.
For the sake of assessing the argument from information processing, I am happy to grant Schwitzgebel's quasi-empirical claim that the levels of information processing in a conscious brain and the United States are similar. What I find problematic is the conclusion that is drawn from this claim. Even if phenomenal consciousness were to be explained in terms of information processing, it is not clear to me that these considerations support group phenomenal consciousness. In fact, this argument strikes me as an instance of a weak analogy: just because a phenomenally conscious creature and the United States are similar to each other with respect to one property (level of information processing), they don't have to be similar to each other with respect to other properties (phenomenal consciousness). This is true even if there is an explanatory connection between information processing and phenomenal consciousness-unless the connection amounts to the claim that high levels of information processing entail phenomenal consciousness. Moreover, even if we accept that phenomenal consciousness can be explained via an information-processing approach, we can still resist Schwitzgebel's conclusion. One approach of this kind is the integrated information theory (IIT) (see Tononi & Koch, 2015). But, as List (2018) demonstrates, group entities do not have phenomenal consciousness even if IIT is true, because the internal structure of many such group entities (like the United States in Schwitzgebel's example) do not have the kind of informational integration that would be needed for phenomenal consciousness according to IIT. 16 Thus, I think the argument from information processing fails also.
To conclude this section, (i) that groups are not phenomenally conscious is the default position regarding these matters, and (ii) the arguments against this default position are wanting. Therefore, I think we have a good case for (A4). Now, putting premises (A1), (A2), (A3) and (A4) together, we get anti-psychologism: propositional attitudes are not mental states.

Objections, Replies, and the Mental Realization Thesis
In this section, I will consider a series of objections to my argument for anti-psychologism, offer replies to them, and elaborate further on how I think propositional attitudes are connected with the mental lives of minded beings.
First of all, one might object to my argument by suggesting that although the individual premises of it may be plausible, there is a problem with how they are put together, or how I motivate them. 17 In particular, recall that I have appealed to functionalism in motivating (A1). That propositional attitudes can be understood as functional properties and that group entities can bear the relevant properties that occupy the relevant causal roles make the thesis of GGI plausible. Now, the worry is that if I am happy to endorse functionalism for propositional attitudes, it is not clear why I shouldn't accept functionalism for phenomenal consciousness and mindedness more generally. For if I endorse functionalism for mindedness more generally, it is dubious that phenomenal consciousness can play the roles in premises (A3) and (A4) I want it to play. More specifically, if functionalism about phenomenal consciousness is true, my rejection of group phenomenal consciousness-hence my case for (A4)may be undermined. After all, if functionalism about phenomenal consciousness is true, phenomenal consciousness can be demystified in the same way that propositional attitudes can be, in which case group phenomenal consciousness may become as plausible as GGI itself. Thus, the objection goes, there is an internal tension in my overall package of views.
I have two replies to this objection. First, it is not obvious that if functionalism is true about the mind (to include phenomenal consciousness), my argument fails. Even if functionalism about phenomenal consciousness is true, phenomenal consciousness can still be the mark of the mental, in which case (A3) can still be defended. Moreover, the truth of functionalism about phenomenal consciousness does not automatically undermine my case for (A4). Even if phenomenally conscious states are functional states characterized in terms of causal roles, this itself does not entail that group entities can have phenomenally conscious states because it is not guaranteed that group entities are (or can be) bearers of the kinds of properties that play the relevant causal roles. Here, a lot hinges on what kind of functionalism would be true and what the relevant functional properties and causal roles would be.
My second reply is that there is a reason why it is admissible not to adopt functionalism for phenomenal consciousness while adopting it for propositional attitudes. As I suggested in my brief discussion of the difficulty of explaining consciousness, views like functionalism are known to face difficulties regarding phenomenal consciousness, because it is not clear if we can explain what it is like to have a conscious experience purely in causal terms. Thus, we can be prepared to be functionalists when it comes to propositional attitudes without having to worry about phenomenal consciousness if we separate these two domains. Indeed, my separate treatment of phenomenal consciousness and propositional attitudes (which is not new; see Chalmers, 1996;Kim, 2005) renders functionalism about propositional attitudes invulnerable to objections from phenomenal consciousness simply because I claim that propositional attitudes do not have much to do with phenomenal consciousness. Recall that the punchline of the objection in question is that there is a tension between holding functionalism about propositional attitudes and rejecting functionalism about mindedness. But this tension is alleviated by treating phenomenal consciousness and propositional attitudes differently to begin with.
A separate, but related objection is that there is an opposing position to my package of views which says that group entities do indeed have propositional attitudes, and that these are mental states of one kind; but group entities are not phenomenally conscious, and thus they lack mental states of another kind. 18 In response, I think there is more agreement than disagreement between this position and my position. Both positions treat propositional attitudes and phenomenally conscious states as importantly different kinds of states. In a sense, the disagreement is almost verbal. But importantly, this is not to say that the entire issue is a verbal dispute. After all, the shared agreement between my position and the position expressed in this objection is in direct opposition to the view that these states are not states of fundamentally different kinds. Importantly, they rule out the position that propositional attitudes are phenomenally conscious states, which takes me to the next objection to my argument for anti-psychologism.
Recall that (A2) and (A3) together entail the claim that only those properties whose instantiations require phenomenal consciousness are mental properties. Now, this next objection to anti-psychologism actually accepts this, and in fact a stronger version of this claim, namely that a property is mental if and only if its instantiation requires phenomenal consciousness. This objection is motivated by Galen Strawson's work, (1994). Strawson's work is especially relevant here, as he explicitly endorses (A2) and (A3) (1994, p. 153), and may accept the stronger claim here; but he would reject the conclusion that propositional attitudes, at least those that are occurrent (i.e., not merely standing, or merely dispositional), are not mental properties. 19 In fact, in Strawson's view, propositional attitudes are phenomenally conscious states. That is, there is something it is like to believe that p; there is something it is like to desire that q; and so on. These claims have recently been defended by a number of philosophers who think that there is cognitive phenomenology, or intentionality and phenomenality are inseparable (see, e.g., Horgan & Tienson, 2002;Pitt, 2004;Kriegel, 2015; see also Bayne & Montague 2011). Now, if there is cognitive phenomenology in the sense that propositional attitudes instantiate phenomenal properties, then anti-psychologism is in trouble: if propositional attitudes are phenomenally conscious states, then the main argument of this paper can't show that they are not mental properties. But I believe that we can rescue anti-psychologism from this objection from cognitive phenomenology by giving an account of the role of propositional attitudes in our mental lives.
I have so far construed anti-psychologism as a negative thesis: propositional attitudes are not mental states/properties. But this doesn't mean that by defending antipsychologism, we can't also hold a positive thesis about propositional attitudes. The relevant positive thesis I have in mind is that propositional attitudes are typically co-instantiated with mental properties, at least in paradigm cases. Cases of GGI are not paradigmatic cases of propositional attitudes, and in such cases, entities that are bearers of propositional attitudes are not also bearers of mental properties. Paradigm cases of propositional attitudes are cases of minded beings like us having beliefs and desires, and in such cases, especially when such beliefs and desires are occurrent (rather than merely dispositional), we also have mental properties in virtue of having phenomenally conscious experiences.
Suppose I have an occurrent belief that it is a sunny day. Presumably, this occurrent belief will be accompanied by a number of mental episodes I will find myself in. For example, I might be visually experiencing clear blue skies and feeling warmth on my skin. I might also say "It's a sunny day!", and hear myself saying these words. These all involve experiences with distinctive phenomenal characters. There is something it is like to see blue skies; there is something it is like to feel warmth; there is something it is like to hear oneself speaking. In this example then, my belief that it is a sunny day is co-instantiated with a number of mental properties corresponding to these phenomenally conscious states.
There are various possible mechanisms and explanations as to why such coinstantiations take place. Some of these mental states are causally related to the belief in question. Plausibly, my visual experience of the clear blue skies and my bodily sensation of warmth causally contribute to the formation of my belief that it is a sunny day, and these experiences may last long enough to co-occur with this belief. Likewise, my belief may cause me to verbally report that it is sunny and my auditory experiences may temporally overlap with my occurrent belief. Although there are various possible mechanisms that can explain such co-instantiation along these lines, a particularly interesting proposal is that we should explain it by means 1 3 Are Propositional Attitudes Mental States? of a realization thesis: my occurrent belief that it is a sunny day is realized by some mental property.
In our current example, the mental property that realizes my occurrent belief is presumably a conjunctive or structural property that includes the aforementioned mental properties (my visually experiencing blue skies, my bodily sensation of warmth, and so on) as its conjuncts or constituents. Here, this mental property realizes my belief that is a sunny day by playing the causal role of this belief. Moreover, this mental property necessitates having phenomenally conscious states (in virtue of being constituted by phenomenally conscious properties), and that is why a conscious experience accompanies this occurrent belief.
Call this the mental realization thesis: occurrent propositional attitudes are sometimes realized by mental properties. If the relationship between my occurrent belief and this conjunctive or structural mental property is realization, it would also be right to spell out anti-psychologism as the thesis that propositional attitudes are multiply realizable by mental properties in some instances and non-mental properties in other instances.
The mental realization thesis aligns perfectly with our choice of theory for understanding propositional attitudes: propositional attitudes are functional properties which are realized by different kinds of properties in different kinds of entities. They are realized by mental properties in minded beings (which are in turn realized by neurophysiological properties), and by non-mental properties in non-minded beings such as groups. 20 The picture we have so far is that my occurrent belief that it is a sunny day is realized by a mental property (which is itself a conjunctive or structural property consisting of mental properties as constituents), and this mental property is realized by neurophysiological properties (in virtue of each constituent property being realized by some neurophysiological property).
In Sect. 2, I mentioned that there are different versions of functionalism. One dimension of differentiating between functionalist theories concerns the relationship between a functional property, its individuating role, and the property that plays that role. According to one version of functionalism, functional properties themselves are the properties that play the causal roles that individuate them. If we endorse this version of functionalism, anti-psychologism will be false for occurrent beliefs, because the mental realization thesis says that the causal roles of occurrent beliefs are played by mental properties. That would entail that occurrent beliefs are mental properties. But there is another version of functionalism which doesn't identify functional properties with the properties that play their causal roles, and this is the kind of functionalism according to which occurrent beliefs can be realized by mental properties without being identified with them. On this version of functionalism, a belief is a functional property, and for something to have a belief is for it to have some property that plays the causal role that individuates that belief. Nothing in this form of functionalism entails that the occurrent belief itself must be identified with the mental property that realizes it. In fact, it is this latter version of functionalism that is strongly associated with the multiple realizability claim that different properties could realize the very same functional property in different systems.
With the mental realization thesis in mind, let's revisit the objection from cognitive phenomenology that propositional attitudes, at least the occurrent ones, are phenomenally conscious states. While explaining how and why propositional attitudes are often co-instantiated with mental properties, the mental realization thesis also explains away the temptation to deem propositional attitudes phenomenally conscious. There has to be a significant difference between (a) believing that it is a sunny day with some or all of the aforementioned phenomenally conscious states and (b) being in the same belief state with none of these (or any other) phenomenally conscious states. In fact, there is nothing it is like to be in (b), while there is something it is like to be in (a). More generally, when a belief is realized by a mental property, there is something it is like for the subject of that belief to be in that state; and when it is not realized by a mental property, there is nothing it is like to be in that state.
If anti-psychologism is true, then it is easy to see why propositional attitudes are never phenomenally conscious. In cases of GGI, group entities are bearers of propositional attitudes, but there is nothing it is like for such entities to have these properties. In paradigmatic cases, such as typical human subjects with rich mental lives, these states are accompanied by phenomenally conscious states, but they are not themselves phenomenally conscious states. From the fact that there is phenomenal consciousness when having propositional attitudes, it doesn't follow that propositional attitudes are phenomenally conscious 21 . The mental realization thesis explains why, typically, there is phenomenology when there are (occurrent) propositional attitudes, and the argument for anti-psychologism shows why there is no phenomenology of propositional attitudes.

Conclusion
I have argued that beliefs, desires and more generally, propositional attitudes are not mental states. My argument is that if propositional attitudes are mental states, then non-minded beings couldn't have propositional attitudes; but some non-minded beings could have propositional attitudes. I have appealed to groups as examples of such non-minded entities.
Before concluding, a final word on a further import of this discussion. Thanks to using groups as examples of such non-minded entities, in addition to making a general claim about the nature of propositional attitudes, the arguments of this paper also have consequences regarding the topic of group intentionality and "group minds". Proponents of group intentionality take themselves to be defenders of the view that there are group minds (e.g. Theiner & O'Connor, 2010); and some opponents of the group minds view reject it based on their opposition to GGI (e.g.,

3
Are Propositional Attitudes Mental States? Baddorf, 2017;Overgaard & Salice, 2021;Rupert, 2004;Wilson, 2004). If the way I have appealed to GGI in my argument for anti-psychologism is right, then we will also need to revise our discussion of the link between GGI and group minds. If I am right, although there may be actual cases of GGI, this doesn't entail that there are group minds.