The unbearable lightness of the personal, explanatory level

I begin this paper by demonstrating that there is a perceived overlap between phenomenology and the personal level. This perception has recently played a decisive role in evaluating phenomenological contributions to discussions within cognitive science, for example, on topics of social cognition. In this paper, I aim not only to understand what might be meant by associating phenomenology with the personal level, but to cast this association in a critical light. I show that the personal level is essentially an explanatory level, whereby perceptions and mental state terms (paradigmatically, belief and desire) explain purposive action. I then separate the notion of consciousness from the notion of the personal level. To do so, I advance Wittgenstein’s private language argument in conjunction with Sellars’ account of how the meaning of mental state terms derives from their explanatory function. Using the Wittgenstinian/Sellarsian picture as guide, I show that characterising personal level explanations by reference to conscious experiences imputes excess baggage over and above the commitment to a unique explanatory level. Yet, for many, ‘phenomenology’ is the level of conscious experience. I argue that it is when the extra baggage of assuming that we are aware of our explanatory, personal level mental states is coupled with the controversial claim that phenomenology is tantamount to the verbalisation of conscious states that the ill begotten association between them is arises.


Introduction
Within philosophy of mind, there are many examples of work that posits an intimate association of phenomenology and the personal level-an association sometimes amounting to identification. Snygg and Combs refer to the "'personal,' or 'phenomenological' frame of reference" as if they are interchangeable (quoted in Ashworth 2006, p. 35). More recently, Di Francesco & Marraffa (2014) mention in passing "the personal sphere of phenomenology" (p. 2) in much the same way as Reynolds (2015) mentions "the typically personal-level descriptions of phenomenology" (pp. 334-335). Herschbach (2008) talks in the same breath about "personal-level, phenomenological experience" (p. 226). Cappuccio & Wheeler (2010) claim outright that "phenomenology is a species of personal-level explanation" (Cappuccio & Wheeler, 2010, p. 131). 1 This association between phenomenology and the personal level is not an idle one. In response to phenomenological critiques of the social cognition accounts of simulation theory (ST) and theory-theory (TT), both Herschbach (2008) and Spaulding (2010) admit that phenomenology provides us with adequate personal level accounts. Still, they both maintain that phenomenological arguments have no 'bite' at the subpersonal level. Moreover, they maintain that neither of the subpersonal versions of TT and ST are committed to any picture on the personal level of phenomenological experience. Accordingly, some authors conclude that the distinction between personal and subpersonal levels is roughly homologous with the distinction between the areas a phenomenologist might contribute to and those they cannot. This pairing goes on to play a significant role in the debate between contemporary phenomenologists and their interlocutors. It therefore seems as if a distinction between the two explanatory levels is mapped onto a distinction between two approaches to cognitive science. This mapping is then leveraged to rebuke phenomenological approaches which are considered to be overstepping their magisterium. With this in mind, it is worthwhile to examine the association of phenomenology with the personal level.
At least sometimes, the thought that phenomenology is somehow tantamount to the personal level arises from two presuppositions which I provide here only provisionally: 1 As the quotes above and below show, the association between the personal level and phenomenology within philosophy of mind is held by psychologists (Snygg and Combs), neuroscientists (di Francesco and Marraffa), cognitive philosophers (Cappuccio, Spaulding, Herschbach), and others who are more directly involved in the hybrid Continental phenomenology/cognitive science/psychology/analytic philosophy space to which this journal is dedicated (Reynolds, Wheeler). I think it is a mistake within the context of this paper to class authors into 'supporters' and 'critics' of phenomenology (because there are a variety of sense of 'phenomenology' (see Sect. 4), and even 'supporters' have their criticisms and vice versa). The association I examine operates during discussions that attempt to limit the scope of phenomenology (Herschbach, Spaulding), and discussions that attempt to advance various positions associated with contemporary phenomenology (see . My paper, therefore, targets the association itself and its often-varied impact on different discussions within philosophy of mind as opposed to any one of these 'camps' employment of it. It is therefore a mistake to read into this paper the advancement or prohibition phenomenological theory. It is merely the attempt to clarify a background distinction and thus the terms of the debate and, once this is done, think through the implications (see . From this, it follows that personal level entities can be decoupled from the level of conscious awareness. Section 3.2 demonstrates that identity theorists would not subscribe to an overlap of the subpersonal with the nonconscious. This gives reason for decoupling the nonconscious from the subpersonal. Section 3.1 and Sect. 3.2 thus show that associating consciousness with the personal is both too wide and too narrow: we are not necessarily aware of personal level states, nor are all subpersonal states nonconscious. Section 4 moves to distinguish between two different uses of the term 'phenomenology' before examining the relationship between phenomenology and the now-clarified personal/subpersonal distinction.

The personal level
I will begin by outlining a definition of the personal level. Some delicacy is required here as there are a variety of accounts of the personal level. The phrase 'personal level' is most often thought to have been given its first and clearest formulation in Daniel Dennett 's Content and Consciousness (1969). In it, Dennett lists Ryle and Wittgenstein as the preeminent personal level theorists that preceded him (Dennett, 1969, p. 95). A perfunctory list of authors who have a theory of the personal level would include Wittgenstein, Ryle, Dennett, Fodor, Davidson, McDowell, Bermudez, and Hornsby. In what follows, I will be careful to outline the personal level by providing only its essential features, and I will avoid awarding undue favour to any particular interpretation (or at least point out when such a favouring occurs; see Sect. 2.3.2). Care should be taken when defining the personal level because, as Drayson (2012Drayson ( , 2014 points out, the notion of the 'personal level' is implicitly packed with baggage. Often the term 'personal' is illegitimately run together with other notions like 'rational', 'deliberative', 'explicit', 'doxastic', and, importantly, 'conscious'. Only once the essential aspects of the personal level are identified and we know the contours of the entity we are dealing with can we advance an adequate assessment of the association between phenomenology and the personal level.

A kernel of agreement
We can define the personal level 4 as a level of explanation. The notion of explanatory levels, though obviously metaphoric, is one commonly employed in philosophy of mind (Craver, 2015). It can be put less metaphorically by just saying that personal level explanations are a sui generis form or type of explanation, a form or type which should be distinguished from other forms or types, and these other forms or types constitute respectively other 'levels' of explanation. The account of personal level explanation lies at the centre of theories of the personal level. Entities that 'exist' at the personal level gain their quiddity through the role they play in a unique style of explanation. As Hornsby (2000) puts it, one "understands the personal level only by allusion to a properly personal style of explanation" (Hornsby, 2000, p. 9). In other words, if we want to understand the way that the term 'personal level' is being used, as say a predicate of certain 'mental state' concepts (i.e., by saying that the concept of 'belief' is a personal level term), then we need to understand the way that those concepts are employed within a certain sui generis type, form, or 'level' of explanation Stating that the personal level is a level of explanation is already informative. On one common construal, all explanations are communicated in language or some other form of semantic entity. 5 On this construal, the constituents of explanations are subject to the kind of success conditions that only semantic entities such as propositions can meet. 6 Despite the array of positions taken about the personal level, it is generally agreed that personal level explanations are expressed in propositions exchanged between persons; any further ontological commitments are debatable (see Sect. 3.1). Thus, even a cursory discussion reveals that, as a form of explanation, the personal level exists on the level of language. This point will prove important to the discussion in Sect. 3 below. also point out (Sect. 2.4), we should also distinguish the philosophical characterisation of the personal level as a sui generis level of explanation from the empirical occurrence of personal level explanations. 5 This linguistic construal may seem too strong for some. It might be objected that some philosophers (i.e., Salmon) take an ontic approach explanation, according to which explanations are non-representational, mind-independent, spatio-temporal entities. Also, for dynamical systems, models count as explanations, and models are not obviously propositions. However, the extent to which dynamical models are nonsemantic is contestable. Moreover, we should distinguish between the weak claim that explanations are semantic from the stronger claim that their explanatory force arises as a result of the logical relations that pertain between sentences. The latter stronger claim is the now-debunked deductive-nomological position. The former weaker claim just entails that, say you have an ontically very real mechanism which causes a phenomenon just out there existing in the world unbeknownst to anyone. Surely no one thinks this ontic entity is yet an explanation. This is not an explanation until an inquirer comes along, wonders about the phenomenon, identifies the mechanism, gains some understanding, and (finally) expresses the ontic situation in a sentence or some other type of communicable vehicle. Surely, the explanatory force derives from the ontic situation (and not, say, the logical relation between the sentences that express it), but the explanation itself is still a semantic entity. Even Salmon seems to endorse this point: for example, when distinguishing the ontic approach from its competitors, he still talks characterises an explanation as the provision of a casual "story" and a causal "account" (Salmon, 1984, p. 297). 6 To say that explanations are semantic is not to deny three other pertinent points. First, that explanations, undoubtedly, have a cognitive cause (a desire for understanding) and a cognitive effect (if successful, the aforementioned state of understanding). Second, explanations occur in an epistemological context which affects the success of the emergence of the cognitive state of understanding. Finally, explanations are, according to some, 'backed' by a metaphysical relationship that exists out there in the world (i.e., causality, grounding, dependence, etc.). However, the cognitive effect of understanding comes about as a result of comprehending certain sentences; the occurrence of an explanation in a certain epistemological context is a linguistic event; it is only once our preferred metaphysical relationship is captured linguistically that they we can have an explanation. See also the previous footnote.
It is natural to wonder at this point which explanations count as personal level explanations. The thing to be explained, as Dennett (1969) puts it, is "people and their… activities" (ibid., p. 93). The explanandum of personal level explanation is, therefore, the chosen courses of action that people take, and thus becomes a component of action theory. The theory of personal level explanation accordingly arose in the context of questions concerning volition. Wittgenstein (1953) formulates the question one way by asking, "What is left over if I subtract the fact that my arm goes up from the fact that I raise my arm?" (Wittgenstein, 1953, p. 161). Personal level theorists seek to provide an account of what would count as an explanation of the broad genus that the latter phenomenon, the raising of my arm, falls under. Personal level explanations are given when we need to explain our own action (to ourselves or more often to another), or when we need to explain another's action to ourselves or to a third party.
Dennett (1969) provides two characterizations of personal level explanation. The first notes that the appropriate personal level explanation of the action of pulling one's hand away from the stove is that it hurt; that there was pain in one's hand (Dennett, 1969, p. 91). Dennett's point is that the personal level is "the explanatory level of people and their sensations and activities" (ibid., p. 93). Thus, one pattern that personal level explanations employ is to provide simple sensation states. In these patterns, sensation states feature as explanantia. However, these sensation states need not be as simple as a pain in the hand; they might instead be complex perceptual states. The point here is that, according to the theory of personal level explanation, (sensory) perceptions explain action. The explanation that 'I ran because I saw a tiger emerging from the bushes' operates at an appropriately personal level and is analogous to Dennett's initial example in the right ways.
However, there is a fuller account we might provide which depends on context and the concerns of the explanatory inquirer. In order to explain the action of fleeing, one might add that one believes tigers to be dangerous and that one wanted to avoid a mauling. Thus, descriptions of perceptions are not the only explanantia that we might adduce to illuminate the nature of action at the personal level. In his second pass at characterising the personal level in Sect. 21 of Consciousness and Content, Dennett (1969) adds that personal level explanations might also "proceed in terms of the needs, desires, intentions and beliefs" of persons (Dennett, 1969, p. 164). A second class of explanantia are variously referred to as 'mental events,' 'propositional attitudes,' 'psychological predicates,' or 'mental state terms/concepts' 7 . Moreover, we can list the sort of processes persons refer to here as well; processes like 'judging,' 'thinking,' 'inferring,' and 'concluding'. As a linguistic level, the personal level is the level at which we employ concepts of a particular type. The point is that they are the sort of psychological or mental concepts that might be attributed to persons. It is the pattern of personal level explanation that employs mental state concepts, and not perceptions, that I am concerned with in this paper.

The distinction between the personal and the subpersonal
Traditionally, we find that the notion of 'the personal level' is often contrasted and distinguished from the subpersonal level -as first proposed by Dennett. As Drayson (2012) notes, "the personal/subpersonal distinction is first and foremost a distinction between kinds of psychological explanation or theory" (p. 8). Put otherwise, 'personal' and 'subpersonal' are terms which denote contrasting semantically evaluable accounts, schemas, or patterns of explanation.

The biological subpersonal
There are two broad points of agreement about the personal/subpersonal distinction to which all theorists subscribe. The first is that the contrast in question is between types of explanations that involve the mental states of persons, and the types of explanations found in the physical sciences. The distinction between two explanatory levels arises not only from an investigation into the conditions that explain action, but also from the philosophical tradition which interrogates the distinction between the sort of explanations we offer in our everyday life and the sort of explanation that employs a different battery of concepts as explanantia -specifically, the sort of concepts defined and employed by naturalised empirical science.
The second point of agreement is that one class of such a battery is the sort of concepts we would find in a purely biological psychology. 'Molecules', 'genes', 'neurotransmitters', 'neuroanatomical regions', 'spike trains,' etc., are all concepts which might be involved in explanation offered at a subpersonal level. When the notion of the personal level emerged in the work of Wittgenstein, Ryle, and Dennett, the subpersonal level was exclusively construed in terms of the concepts of a biological psychology. Dennett (1969) himself originally held that subpersonal explanation proceeds mainly by reference to the "brain and events in the nervous system" (Dennett, 1969, p. 93).

The cognitive subpersonal
However, one idea that some contemporary theorists subscribe to is that the subpersonal level of explanation can really be divided into two distinct semi-levels. The lower semi-level is composed of the aforementioned biological concepts, and the upper subpersonal semi-level is composed of cognitive concepts 8 . Di Francesco & Marraffa (2014) characterize this cognitive semi-level as "the information-process-ing level, wedged between the personal sphere of phenomenology and the subpersonal domain of neurobiological events" (2014, p. 2). In at least traditional cognitive science, the cognitive semi-level is populated by concepts which denote functional or computational processes. Subpersonal explanations commonly involve functional analysis that break a system up into subfunctions or subcapacities (Cummins, 1983). The cognitive subpersonal semi-level is not without controversy. As Drayson explains, much like "the overall capacity, each sub-capacity is specified in psychological terms, and so each subsystem is like a 'subperson' who discriminates, evaluates, calculates, remembers, or suchlike; hence the term 'subpersonal psychology'" (Drayson, 2014, p. 339). Thus, the sort of concepts used to describe processes at the cognitive semi-level are the same (at least nominally) as the psychological concepts employed at the personal level. One common complaint here is that subpersonal psychology commits a mereological fallacy in attempting to ascribe properties (i.e., mental states) to parts (i.e., homunculi or neurons), which can only be properly ascribed to the whole person. This is why the Wittgenstinian/Rylean tradition prefers to talk only about a biological subpersonal level (see Drayson 2012, p. 3). One way Dennett avoids the mereological fallacy is to suggest, after Consciousness and Content, that the ascription of psychological concepts at the subpersonal level should not be taken literally. Subpersonal psychology is carried out by the method of homuncular decompositional analysis coupled with a thoroughgoing instrumentalism about psychological explanation. Homunculi -that is, the instrumental embodiment of functional processes -that think or believe are akin to the equator or centres of gravity. These concepts merely appear as useful or explanatory fictions or heuristics (Dennett, 1987).

Is this characterization outdated?
Here, I will briefly address a critical objection to which my characterization of the personal/subpersonal distinction might be subject. The variety of mental state terms we might employ at the personal level have historically been reduced to broad proto categories of belief, desire, and their amalgamation into a state of intention. Recent scholarship denies this coarse-grained, cognitivist framework. There are two claims that have emerged that serve as an objection to my characterization of the personal level: 1) The types of social explanations that people do offer rarely explicitly refer to beliefs and desires (Goldy, 2007;Ratcliffe, 2007), and. 2) As far as our social life goes, people rarely engage in the practice of social explanation at all (Gallagher & Hutto, 2008).
There are two commensurable strategies one might use to respond to these objections.
The first is to stress the distinction between philosophical and empirical contexts. In the context of philosophy of mind, one might respond by saying that it is undoubtedly the case that the explanatory schema I have just outlined has always been the one discussed under the rubric of 'the personal level'. To outline different schemas of explanation such as the ones that Goldy and Ratcliffe emphasise represents a change of topic. Perhaps these newer schemas shouldn't really be deemed 'personal level' explanations. Rather, they might be better considered under the banner of some other type of social explanation. The points of Goldy and Ratcliffe might be better characterized as a challenge to a certain account of versions of folk-psychological explanation and not a challenge about what the personal level consists of per se. Philosophical discussions of the personal/subpersonal distinction include questions over the source of explanatory force. Personal level explanations might be nomological (as Ryle (1949, p. 76) and Fodor (1991) thought), conceptual (as Ryle (1949, p. 71) thought), rational (as Davidson (1980) thought), or causal (as Fodor (1991) thought). Others have questioned whether this explanatory schema differs in kind from those employed in natural science (particularly in biological/subpersonal psychology), and how these two explanatory schemas relate (see Colombo 2013; Davies, 2000;Williams, 2020a). These philosophical discussions might continue even if we agreed that points one and two were true. They might continue just as one discusses the features of nomological mathematical explanations and compares them to scientific explanations, without holding that mathematical explanations are common/pervasive, or the only form of non-causal explanation on the market. One might, after all, be a pluralist about non-scientific explanations and interested in all of their different possible forms.
However, even if one accepts these two points, they risk making the discussion I provide here purely academic. The other strategy I employ in responding to these objections is thus to shift into the empirical register and point out that the above-mentioned critiques of social explanation seem to overlook empirical evidence suggesting that we engage in social explanation with a non-negligible frequency. They also overlook the evidence that such explanations operate by the attribution of concepts drawn from the personal level framework of beliefs, desires, and intentions.
Explanations are answers to questions. Personal explanations arise in contexts where people seek to understand why an action occurred. Broadly, any circumstance where we find ourselves wondering about an action taken by either ourselves or another represents ground for personal explanation. Contrary to the claim that we are rarely at a loss to understand the action of others, in one study 28 participants found themselves 'wondering' about the action of others a total of 239 times over the course of four days, and such 'wonderings' are the cognitive predecessor of personal level explanation (Malle & Knobe, 1997b). This amounts to more than twice a day per person. Even Gallagher & Hutto (2008) admit that there are "puzzling cases" of human behaviour that might cry out for explanation; their point is "simply that most of our everyday interactions are not of this sort" (Gallagher & Hutto, 2008, p. 19). This is surely so; two events per day represents but a fraction of our social engagement. Nevertheless, if we leave aside the question of whether our primary engagement with others is along explanatory lines, there is at least some evidence that the grounds for personal level explanation arise with enough frequency to save the present discussion from becoming purely academic.
Moreover, the empirical version of the theory of personal level explanation is that people satiate the explanatory hunger which arises from such wonderings via the process of attributing concepts of 'belief,' 'desire,' and 'intention' to agents. There is research that suggests this theory is correct. First, studies show that the concept of 'intention' is a concept that 'the folk' really do employ, not just a philosophical notion (Malle & Knobe, 1997a). When asked to define the concept of intention, 96% of respondents in one study employed the key psychological categories of belief and desire. Other evidence suggests that over 90% of explanations of intentional action involve the attribution of beliefs and desires (Malle, 1999).
Taken together, these empirical studies of explanation represent an empirical response to the charge that personal level explanations are neither common nor operate on the belief/desire paradigm. It is surprising that these studies are rarely engaged in recent critiques. They are not discussed in Ratcliffe's whole 2007 monograph which critiques belief/desire explanations; and they are not cited in the papers by Goldy (2007) or Hutto and Gallagher (2008). Naturally, as with all empirical research, only further engagement could deem these issues decided. These disputes will not be the focus of this paper. I merely refer to them here as grounds to defend against the potential objection that my characterization of the personal level represents an outdated and disproven model of explanation that is irrelevant to any empirical discussion of psychological and social understanding.

Uncoupling the levels of the Personal and the conscious
With the essential (or consensus) characterization of the personal level outlined, I will begin to remove philosophical baggage that proves to be of pivotal importance to the association of phenomenology and the personal level. I mention above that the notion of the personal level is often conflated with other notions. One of the most persistent conflations is of the personal and the conscious, and of the subpersonal and the nonconscious. This conflation is commonplace. Frankish (2009) states that we "are typically conscious of our personal mental processes" (p. 92). Davies (2005) talks in the one breath about the "personal-level events of conscious thought" (p. 370). Conversely, the subpersonal is routinely assumed to be the realm of the non-conscious. Burge (2003) characterizes the subpersonal as "not accessible to introspective or reflective consciousness" (p. 384). The motivation for this conflation might be the thought that we are conscious of the referents of terms employed in personal level explanations, and that subpersonal mental processes are, by their nature, inaccessible by consciousness.
I will show that the connection between conscious awareness and the personal level is contingent at best. This connection hinges on one's theoretical orientation, and it is almost certainly the case that some personal level mental states are not attended to by consciousness. One might be committed to the above account of personal level explanation and have only very minimal commitments to consciousness.
In what follows, I show that there are a series of positions which do not conflate the notions of consciousness and the personal level. However, I stop short of endorsing any of these positions; their import becomes clear in Sect. 4.

The position space on the metaphysics and semantics of the mind and explanation
One way of thinking that opens the possibility that we are not conscious of certain personal level mental states stems from the linguistic nature of the personal level itself. As a preliminary point, merely recognising the validity of the personal level of explanation does not commit one to any theory concerning how explanatory terms refer, or to what they refer, metaphysically speaking. There are a variety of wellknown metaphysical positions on the referents of mental state terms. One might assume our mental state terms refer to brain states (Feigl, 1967), functional states (Lewis, 1966), and one might be sceptical about whether mental state terms refer to anything at all (Churchland, 1984). However, considering that one might be purely coherentist or instrumentalist about meaning and explanation generally, one could occupy any of these metaphysical positions without altering one's account of how personal level terms fulfil their explanatory function (see Drayson 2014 for a fuller discussion).

Wittgenstein and Sellars
Regarding the question of mental state semantics, one alluring theory holds that the meaning of mental state terms arises by the very role those terms play in explanatory practices. To arrive at this theory, one begins by applying Wittgenstein's (1953) argument against private language to mental states and derives the thesis that the meaning of mental state terms is established via public practices (not private experience). An important detail is that the meaning of mental state terms is established through the use or function of words. At least on the behaviourist interpretation of Wittgenstein, mental state terms function as proxies for descriptions of intelligent behaviour.
But one might equally suspect that mental state terms gain their meaning precisely because they explain intelligent behaviour, not because they describe this behaviour. The important shift here is to think of mental state terms as types of theoretical or explanatory concepts, not as descriptive concepts. To detail this basic picture, I draw on the work of Wilfrid Sellars, especially his Empiricism and the philosophy of mind (hereafter EPM 9 ).

Sellars functional role semantics and the myth of genius Jones 10
Sellars accepts Wittgenstein's claim that the meaning of items in our language is established by the functional role those terms play in pattern governed linguistic exchanges. When I say, for example, that 'term x means y', I am saying that term x is the sort of thing that plays functional role y. Here, a functional role is a reliable 9 Hereafter, I follow standard practice in Sellarsian scholarship in providing section and paragraph numbers (i.e., 'EPM, XIII.53') for EPM, not page numbers. 10 Even though, as Drayson (2014) notes, Sellars' distinction between the space of reasons and the space of causes is often mapped by the Pittsburgh school onto the personal/subpersonal distinction, I draw on him here to highlight the reorientation in thinking that he demands concerning a type of explanation that is for all intents and purposes identical to personal level explanations pattern of use consisting of three essential types of pattern-governed linguistic behaviour: appropriate responses to perceptions (i.e., 'entry'), inferential intra-linguistic transitions, and volitional language (i.e., 'exit'). On this account, an individual knows the meaning of the word 'red' when they reliably use phrases like 'this is red' in the presence of red objects (entry), are disposed to utter correct inferences about red things, and reliably behave in ways commensurate with their own utterances about red things (exit) (Sellars, 1980, p. 69). Thus, the meaning of any term is determined by "the uniformities that characterize the learned, holistic, pattern governed linguistic behaviors of speakers of that language" (O'shea, 2007, p. 61).
Given the public and intersubjective emphasis of Sellars' account of meaning, one aim of the infamous myth of genius Jones in EPM is to suggest how the idea that mental state terms refer to inner events that carry with them a private phenomenology could have developed independently from introspective or reflective access. Sellars grants a mythic community -the Ryleans -the resources of.
(i) a vocabulary for public behaviour, (ii) a working functional role semantics that affords them the concept of 'intentionality' and other semantic concepts (EPM, XII.49), and.
(iii) a theoretical discourse that allows them to construct crude explanations about unobservable causes that differ in kind from observable events (EPM, XII.52).What this mythic community denotes when they use mental concepts is merely a Rylean disposition to verbally 'think out loud.' What baffles the Ryleans and cries out for explanation is silent, rational action. Imagine a member of this community, genius Jones, in line for a train ticket with other members of his community. Suppose that Jones observes Sally silently fidgeting, looking around, and then suddenly switching to another ticket line. Jones might wonder about Sally's behaviour, and in response to his wondering, posits that something is going on in Sally which is in its semantic dimensions similar to, and which she would naturally have expressed by muttering to herself: 'Hey, that line is moving faster than this one'. The cause of her silent line-switching behavior, the mythic genius Jones now proudly teaches his fellow Ryleans, was the occurrence in her mind of a silent inner thought, a thought that is language-like in (at least) the sense that it has the same meaning or propositional content as the sentence Sally would have mumbled to herself had she been in a thinking-out-loud frame of mind. (O'Shea, 2012, p. 186) Thus, the "postulation of inner functional roleplayers in a kind of 'Mentalese' is part of a proposed causal explanation as to why those behavioral patterns are the way that they are" (O'shea, 2007, p. 93).
In proto-scientific fashion (EPM, XIII.51.2), Jones makes an explanatory inference and attributes to Sally a new type of unobservable, theoretical entity -a thought. This thought has the same content and plays a role that the linguistic community would approve of were Sally's behaviour accompanied by an attendant commentary. In this way, language serves as the known domain on which these new types of theoretical entities are modelled (EPM, XIII.51.1). Motivated by a familiarity with other types of proto-scientific explanation stipulating that non-observable, inner-goings-on are the cause of observable goings-on, genius Jones identifies a distinctly inner type of mental state as the cause of Sally's behaviour (EPM, XV.56). The key features of mental states (i.e., 'inner events' (EPM, XV.58.2), which are intentional in nature and have semantic content (EPM, XV.57) that cause purposive action (EPM, XV.56)), is an effect of the nature of our explanatory hypothesis. The upshot for Sellars is that we cannot rule out the possibility that the nature of our psychological vocabulary as it is used today has its genesis in the intersubjective speech-act of explaining the silent actions of our co-community members by personal level explanation.
Before moving on, I will highlight some features of Sellars' account which make it more suitable for introduction to the context of a paper about phenomenology. First, the Sellarsian picture is a philosophical theory about meaning, specifically about how certain uses of language reveal and parse reality. It is not a theory about how we invent mental reality. As DeVries & Triplett (2000) put it, "Jones's postulation of thoughts is part of the self-discovery of reason, and reason cannot 'undiscover' itself once it has come to self-consciousness" (DeVries & Triplett, 2000, p. 135).
The conceptual priority and theoretical nature of mental language does not afford the possibility that we might radically revise or eliminate our mental states any more than we might eliminate other theoretical entities such as black holes (EPM, XV.58.2). Thus, we ought to stop short of thinking that Sellars' account implies a possible eliminativism about mental states (a topic he disagreed with his PhD student Paul Churchland about).
Sellars is in the unique position of balancing, on the one hand, the conceptually mediated nature of our access to our own thoughts (i.e., avoid the form of the myth of the cognitive given) with, on the other hand, the aim to do philosophical justice to cognitive phenomenology (see Williams, 2021b andSellars 1981). His psychological nominalism (EPM, XI.29) merely implies that inner mental life can only be known and referred to after we have formed an intersubjective, inferentially integrated conceptual vocabulary about it (EPM, XVI.62). Sellars' account shows that we can conceptually uncouple the notion of conscious awareness of the mental from explanatory acts that refer to the mental, and that perhaps explanatory acts precede consciousness in the order of conceptual priority vis a vis mental state terms.
Second, Sellars' account is not an empirical theory about how we actually socially cognise à la TT. One of the effects of EPM is an erosion of the theory/observation distinction. Thus, any post-Rylean society might very well have integrated the conceptual framework of genius Jones such that intersubjective perception is conceptually laden to the point where we can reliably and non-inferentially 'observe' mental states 'in' the action of others. This would be loosely akin to how cloud tracks in a cloud chamber count as an 'observation' of ionised particles for the trained scientist (see Brandom 2002;cf. O'Shea, 2012). One can only juxtapose the Sellarsian account with contemporary social cognition discourse. This is because he does not seem to have anything to say (in EPM at least) about the potential empirical neural/functional, organisational underpinnings or the phenomenology of contemporary social cognition. Sellars' mythological fiction is an attempt to develop a philosophical account about how we might have developed an inferentially-integrated vocabulary for mental life in the first place; a theory that, in the style of a Kantian synthesis, does not resort to the philosophical oddities of an austere logical behaviourism nor resurrect the ghost in the machine that Ryle has exorcised (DeVries & Triplett, 2000, p. 128). To accomplish this goal, the Sellarsian picture postulates a genetic relation between mental state concepts and personal level explanations.

The import of the Sellarsian Picture for Accounts of the personal level
As I mentioned at the outset of this section, I am not looking to endorse this picture wholesale. For the purpose of this paper, it is just important to recognise that there is at least one way to view the situation such that it is not that our philosophical account of mental states serves to enlighten us as to the nature of the sort of terms we employ during personal level explanations, but the other way around. Noting the importance of personal level, social explanatory practices provide a loadbearing steppingstone in our philosophical account of mental states. The philosophical account which results claims that mental state terms have earned their place in our public vocabulary precisely because the notion of an internal mental state does some work in the practice of explaining the action of persons. At least, this is true for Sellars. This is the startling idea rooted in the rather prosaic observation that the personal level is an explanatory, linguistic level that provides mental state terms as explanantia. Perhaps the very meaning of the terminology of the personal level derives from the very role those terms play within those explanations.
When they propose that mental state terms are defined, in part, by the role they play in public linguistic practices like explanation, Wittgenstein and Sellars dispel the notion that the meaning of mental state terms is tantamount to privately accessible conscious states. We are being drawn away from the picture which holds that when we use mental state terms, we are pointing backwards towards something inside the Cartesian theatre located inside our skulls by an act of interior ostension. Instead, this view points outwards towards practice of explaining action as the source of meaning. As I say above, it is not that Sellars thinks we are non-conscious philosophical zombies. Sellars just aims to point out that we cannot dismiss the possibility that the terms involved in personal level explanations have an explanatory function before they take on a secondary and conceptually separable reporting role (see EPM, XV.59). The attribution of mental states to others has conceptual priority over the capacity to characterize self-experience in mental terms. Therefore, emphasising the explanatory nature of the personal level serves to provide a philosophical account which respects both the intersubjective nature of meaning and our reflective awareness of cognitive phenomenology, as opposed to taking the latter as brute and unmediated fact. In doing so, the philosophic account affords us the possibility to consider explanatory acts divorced from the phenomenal context in which they are seemingly prima facie found.
To conflate the personal level and consciousness, or to assume that it is in the nature of personal level postulates that we are aware of them, is to fail to appreciate the variety of possible philosophical positions on personal level explanations. Moreover, the theory of personal level explanation, on some influential renderings, is one which fits within a network of theses designed to show how our mental state concepts gain their meaning and play the role they do within language whilst only loosely cou-pling them with the concept of 'consciousness' and 'first-person awareness'. Talking about the personal level in the same breath as we talk about the level of explicit or reflective consciousness is too simplistic.
Of course, it has been argued that we cannot provide a coherent account of features of cognitive states like intentionality without referring to their conscious aspects (Strawson, 2004). Others suggest that it is a categorical mistake to give an account of cognitive concepts such as 'theorising' without making reference to "processes that involve reflective consciousness" (Gallagher, 2005, p. 215). On the other hand, Braddon Mitchel and Jackson hold that cognitive states like belief and desire are considered "prime examples of states for which there is not something it is like to be in them," i.e., states which lack any phenomenal consciousness (Braddon-Mitchell & Jackson, 1996, p. 295;quoted in Montague 2016, p. 25). I am not arguing that all personal level states lack an essentially phenomenological component, only that some can, and I think that these claims need to be considered on a case-by-case basis (see 4.2 below). All we need recognise is that the indispensable role that personal level states play in personal level explanations is one good reason to countenance that the connection between some personal level explanatory postulates and consciousness of them is loose and intermittent (for discussion see Smithies 2013), and we should not just associate the conscious and the personal tout court.

Some subpersonal states are conscious
Having shown that some personal level states are potentially nonconscious, I will now show that some subpersonal states can be conscious. Drayson (2014) subscribes to the thesis that we cannot use the personal/subpersonal distinction as a proxy for the distinction between conscious and nonconscious, but for reasons inverse to the ones I have just outlined. She begins by observing that the relationship between the posits of personal level explanations and the posits of subpersonal explanations depends on metaphysical commitments about the mind. One might be committed to a functional identity theory. In order to provide an integrated, multilevel, vertical, functional explanation, homuncular functionalism begins with by identifying everyday mental states like belief with functionally characterized, subpersonal states of affairs.
As one decomposes these functional states by analysis into increasingly simpler levels, subpersonal functional states which have no personal level counterparts quickly emerge. The functional language that one employs to characterize these further states might still employ personal level terminology (i.e., one might say that what these subpersonal states do is 'infer', 'calculate', etc.) but only instrumentally so without engendering a commitment to all the associated features that other personal level states have. But the functionally-characterized subpersonal posits we began with are just personal level states under a different description that do carry with them attribute commitments.
This strategy of analysis arises not only for the functionalist, but for any theorist who holds conceptual/explanatory dualism together with ontological monism; at some stage, states which are metaphysically identical will need to be characterised in differing conceptual vocabularies but the ontological features will be shared. This is the fact that Davidson leveraged to argue that we can maintain that personal level explanations are sui generis but that reasons have causal properties (Davidson, 1980). This implies that some subpersonal states (the ones that are identical to some personal states) have the same attributes as personal level states (assuming these attributes really do metaphysically inhere in the states and do not arise merely from altering linguistic or epistemological context). For example, some mental states are doxastic, i.e., inferentially integrated and consciously accessible 11 . Drayson (2014) notes that Stitch holds that doxastic states "are not only the posits of personal psychology, but also the posits of subpersonal psychology" (p. 343). Thus, Stich's distinction between doxastic and subdoxastic does not divide those states posited by personal explanations from those posited by subpersonal explanation. Instead, doxastic states appear in both subpersonal and personal explanations, whereas subdoxastic states appear only in subpersonal explanation (ibid.). The point here is that some of the mental states that we refer to when we give personal level explanations are the same states found in subpersonal explanations. These explanations carrying with them all their personal level features. At least, this is true for some theoretical orientations.
I suggest that we can transpose this discussion over to conscious states. If one is an identity theorist who thinks that redescribed personal level states feature in subpersonal explanations, and if one further thinks that at least some personal level mental states are sometimes consciousness 12 , then one is obliged to ascribe consciousness to our subpersonal entities as well.
Identity claims underpin much contemporary cognitive science. Cognitive psychologists generally avoid suggesting that they are not discussing the same thing we are when we employ a personal level framework. They merely talk about other things besides, and transpose personal level states into explanatory frameworks (like functionalism) which afford integration with naturalised empiricism (many of these ideas are explored with reference to phenomenology in the recent work by Pokropski 2021). Drayson (2012) notes that a "common misinterpretation of the personal/subpersonal distinction involves taking it to distinguish what is conscious from what is unconscious" (Drayson, 2012, p. 15). She further notes that if "we simply want to distinguish… between conscious and non-conscious states, we can do this without using the terminology of the personal/subpersonal distinction" (Drayson, 2014, p. 344). Associating the personal level with consciousness is both too wide and too narrow. It is not necessary that personal level states have the property of consciousness, and the consciousness of a state is not sufficient to rule out it being a subpersonal state.

Phenomenology and the personal level
I will conclude by discussing the association of phenomenology and the personal level. Here, we might ask what we mean when we refer to 'phenomenology'. I have thus far only given a provisional characterisation of phenomenology (Sect. 1), and it is time to question it. One line of thought holds the concept of 'phenomenology' is essentially tied to the articulation of mental states that we are conscious of, or the articulation of those aspects of mental states that are conscious. Related thinking holds that the term 'phenomenology' is somehow related to the 'what-it-is-like' of mental states ('what-it-is-like' being synonymous with conscious properties) 13 . When used in this way, 'the phenomenology' of such-and-such is constituted by those properties that make up our conscious experience of such-and-such. Carrying out a phenomenological analysis thereby involves reporting or describing these properties. This is what I have referred to in other papers as the 'analytic' (or 'little p') sense of phenomenology (Williams, 2020a). I am not interested in arguing in this paper about the 'authenticity' or 'validity' of this use of the term 'phenomenology' (see Williams, 2021a;cf. Zahavi, 2018). In the present section I will provisionally accept that this is how the phrase 'phenomenology' "is standardly used in present day [analytic] philosophy of mind" (Montague, 2016, p. 8). I am interested in drawing out the implications of thinking of 'phenomenology' in this way with regard to this paper so as to bring some clarity to the ongoing discussion between contemporary phenomenology and cognitive science. Thus, I will accept an appropriately modified version of P1 from the introduction: P1) Little-p phenomenology involves the verbalisation of conscious mental states. Given this limitation and definition, I will make two comments that aim of clarify the relationship between 'p'henomenology and the personal level.

The subpersonal and little-p phenomenology
First, one interesting conclusion (one that ought to be considered by Herschbach and Spaulding) is that talk about subpersonal phenomenology is not incoherent, especially given what I have said above about some subpersonal states being conscious. This will hinge upon one's commitments regarding the relation between personal and subpersonal states. If the initial link in the vertical explanatory chain that subpersonal cognitive psychology forges involves merely re-characterizing or analyzing personal level states into a different and subpersonal framework, then these subpersonal states will have an associated phenomenology, and so we can indeed talk about 'the phenomenology' of some subpersonal states. In summary, merely showing that an account is subpersonal is not tantamount to barring the phenomenologist from contributing to the discussion.

The personal and little-p phenomenology
It is occasionally simply assumed that determining that a process or state is a personal level one brings it within the domain of little-p phenomenological inquiry. I think this assumption is a plausible explanation for the association of phenomenology and the personal level that is found in the quotations from Herschbach (2008) and Di Francesco & Marraffa (2014) that I include at the outset. However, given our uncoupling of consciousness from the personal level, the phenomenologist, considered as someone who verbalises conscious states, cannot always even provide a personal level explanation. Moreover, they are not necessarily in an authoritative position to provide an account of the personal level. Though phenomenology in the little-p sense might provide some information about the subset of personal level states that we are sometimes conscious of, it will not provide information about the entire set of personal level postulates unless you maintain an untenable adherence to a parallelism between explanatory postulates and cognitive phenomenology.
It is not enough to merely show that a term is at the 'personal level', or one that we might make use of in our everyday explanations, in order to bring it within the little-p phenomenologist's purview. Thus, any 'phenomenological' examination and critique of any account of the personal level hinges on prior demonstration that the states referred to by the account carry with them an associated cognitive phenomenology. This is assuming, of course, that the phenomenological critique relies on our intuitions about what we consciously experience.
Gallagher and Zahavi's (2008) attack on personal level versions of TT and ST, for example, would require this prior demonstration as they appeal to "common experience" (Gallagher & Zahavi, 2008, p. 196), and to the fact that the terms which theorytheorists use (terms like 'prediction' and even 'explanation' itself) seem to imply some degree of reflective consciousness "in our everyday psychology" (Gallagher, 2005, p. 215). However, to prove that some term ought to be thought of as implying consciousness, it will not suffice to merely show that the aforementioned terms are personal level terms. More argument is needed to make this case.
I think Gallagher has this style of argument in mind when he states that the vocabulary of theory-theory implies conscious, explicit processes. A similar effort is made in Gallagher & Hutto (2008). In this work, they claim that "neurons cannot pretendthey either fire or they don't", and they claim that as 'pretence' or 'pretend' is one of the key processes involved in simulation, "simulation, as defined by ST, is a personallevel concept that cannot be legitimately applied to subpersonal processes" (Gallagher & Hutto, 2008, p. 20). Aside from the criticism that I have been urging in this paper, i.e., that merely showing a term to to be personal is not enough to show that it involves consciousness, it might also be replied that the subpersonal level accounts of theorising and simulation rely on terms that gain their quiddity owing to their explanatory usefulness. Claims such as Gallagher's misunderstand the strategy of explanatory instrumentalism. Subpersonal explanations frequently redeploy personal level terms if there is an explanatory payoff, even ones that do involve conscious processes. Moreover, even if one thinks, as Wittgenstein and Ryle did, that it is not good practice to describe neuronal processes in personal level terms, one might just as well think we are really talking about the information processing subpersonal level. This level spans from complex processes at the top of the analysis down to basic processes instantiated on neural hardware, and each sub-level is itself characterized in terms frequently found at the personal level.

Whither Phenomenology?
Many readers will surely object to the 'little p' characterisation of phenomenology, and I have sympathies for their complaints. If one returns now to the series of quotations from the beginning of this paper, it becomes clear that not all those authors were operating with the so-called 'little p' sense of phenomenology. Many of them are learned proponents of 'Big P' Phenomenology. This is represented by philosophical phenomenology in the tradition of Husserl. Accordingly, there is certainly another motivation regarding the association of phenomenology and the personal level here; motivation that concerns deeper, perceived connections between some proponents of the philosophy of the personal level and philosophical phenomenology. These deeper philosophical connections have already been the topic of one paper (Williams, 2020b) and will feature as the subject of another paper in the future. Determining exactly what was imputed by authors who draw a connection between big-P phenomenology and 'the personal' (see footnote 4) possibly requires more speculation than is desirable, and certainly more space than is left here.

Conflict of interest
The author declares they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.