1 Introduction: Do Consciousness and Self-Consciousness Go Hand in Hand?

For a long time, the problem of experience and the problem of self-consciousness were discussed separately. The problem of experience was taken to be the “hard problem” of consciousness (Chalmers 1996), deserving the most attention, while the problem of self-consciousness was seen as a distinct and comparably easy problem to solve. This has changed in recent years. Today, a growing number of philosophers argue for an essential connection, a necessary link between experience and self-consciousness. They claim that each and every experience involves self-consciousness, in a way yet to be specified. I call this claim the universality claim or just universalism,Footnote 1 and its proponents universalists. Opponents of the universality claim, i.e., non-universalists, by contrast, hold that there are at least some experiences that do not involve self-consciousness. Typically, they assume that, in many or most cases, experience and self-consciousness go hand in hand. Sometimes, however, according to non-universalism, there is experience without self-consciousness.

This paper is dedicated to the debate between these two camps. More particularly, it is dedicated to what Zahavi, as the leading contemporary universalist, calls the “central argument” (Zahavi 2018, p. 711) for the universality claim. This argument, which he describes as “a kind of transcendental argument” (Zahavi 2018, p. 711), calls into question certain reductive explanations of self-consciousness by demonstrating that they have to presuppose a kind of self-consciousness that they cannot reduce. And therefore, the explanations fail.

I am sympathetic to the transcendental argument. I take it to constitute a forceful objection to the mentioned explanations. I do not believe, however, that the argument leads to universalism in any straightforward way. Pace Zahavi, I argue that the argument cannot vindicate the universality claim. I point out the limitations of the transcendental argument by way of three main objections. And I contend that, even if successful against certain reductive explanations of self-consciousness, the argument simply does not imply the universality claim. Since the transcendental argument appears to be the main motivation for Zahavi’s universalism, I conclude that his position is insufficiently warranted. Given the difficulties outlined, it does not have the argumentative support that seems needed.

2 Defining the Terms

When it comes to defining the terms used in the universality claim, the definition of ‘consciousness’ seems to be the easy part: ‘Consciousness’, as it is used in the debate on universalism, stands for ‘phenomenal consciousness’. Conscious states are taken to be states that are phenomenally conscious. And phenomenal consciousness is characterized by way of Nagel’s classical formulation: Phenomenally conscious states are states such that there is something it is like to have them. They are characterized in terms of their what-it’s-like-ness (Nagel 1974).

I use the terms ‘phenomenally conscious state’ and ‘experience’ synonymously. ‘Experience’, thus, stands for ‘conscious experience’. This raises the question of whether experiences are essentially conscious or whether there can also be non-conscious experiences. The possibility of such non-conscious experiences has been defended by some proponents of higher-order theories of consciousness. According to them, experiences do not have to be conscious. They rather acquire consciousness when a higher-order mental state is directed at them. Rosenthal, e.g., speaks of non-conscious experiences, which, as a result, do not have any what-it’s-like-ness: Since they are below the threshold of consciousness, there is nothing it is like to undergo these experiences (Rosenthal 2005; see also Carruthers 2000). Rosenthal even assumes a kind of phenomenality which is not conscious. In his view, not even phenomenal character is essentially connected to consciousness (Rosenthal 2005).

In what follows, I will remain neutral with regard to the question of whether phenomenal states and experiences can be without consciousness. This question does not need to be answered for present purposes. For even if there are phenomenal states and experiences which are not conscious, the universality claim is not aimed at them. Universalism, rather, is a claim about conscious phenomenal states and experiences. Its scope is restricted to phenomenal states and experiences which are accompanied by consciousness. They are my sole focus in what follows. I will be concerned only with conscious phenomenal states and experiences.

Also, I will not address the question of which states, exactly, make up phenomenal consciousness: Is it limited to sensory states such as perceptions and raw feels? Or does it include cognitive states such as thoughts and beliefs as well, as proponents of cognitive phenomenology claim?Footnote 2 We do not need to decide. What universalists are committed to is the claim that if a state – be it a perception, a thought, etc. – is phenomenally conscious, then this state essentially involves self-consciousness. The question of whether a particular state is phenomenally conscious can be left open.

Now, after this definition of the term ‘consciousness’, as used in the universality claim, we come to what seems to be the harder task, namely the definition of the second term used in this claim, i.e., ‘self-consciousness’. The definition of this term appears to be harder because it is deployed in strikingly different ways by different universalists. There is no commonly accepted definition of ‘self-consciousness’ in the universalist camp. Sartre, e.g., in The Transcendence of the Ego, goes so far as to claim that self-consciousness, in its most basic form, which he calls ‘pre-reflective’, does not involve any self, ego, or person whatsoever. According to him, pre-reflective self-consciousness is anonymous, non-egological, and impersonal. The term ‘self’ in ‘self-consciousness’, thus, merely has a reflexive function: It denotes that what we are dealing with is the consciousness that an anonymous, non-egological, and impersonal consciousness has of itself (Sartre 1957).Footnote 3

Most universalists today hold a different view of self-consciousness. They claim that self-consciousness is not anonymous. It rather involves a self. There is, however, no agreement among universalists as to what this self is, ontologically speaking. Universalist views on the ontology of the self differ significantly. There are physicalist as well as dualist proposals, accounts of the self as a substance, a process, and even, it seems, as a property of phenomenally conscious states.Footnote 4 In what follows, I will leave questions about the ontology of the self aside. In my definition of the universality claim, I will simply take for granted that some kind of self exists and that self-consciousness somehow involves this self. The question, then, is: How? What role does the self play in self-consciousness, according to universalists? In a passage frequently cited to introduce the universality claim,Footnote 5 Zahavi answers this question in the following way:

One commonality [of experiences] is the quality of mineness, the fact that the experiences are characterized by first-personal givenness. That is, the experience is given (at least tacitly) as my experience, as an experience I am undergoing or living through. Given this outlook, it is natural to argue that self-awareness is of pertinence for an understanding of phenomenal consciousness. In fact, phenomenal consciousness must be interpreted precisely as entailing a minimal or thin form of self-awareness. On this account, any experience that lacks self-awareness is nonconscious […]. (Zahavi 2005, p. 16)

Zahavi, thus, defines self-consciousness in terms of mineness, and he defines mineness as my experiences’ being tacitly given to me as my experiences, as experiences that I undergo. Self-consciousness, hence, involves a self in the following way: Experiences are essentially given to selves as their experiences. They are essentially given to selves as belonging to them.Footnote 6

In view of this, the universality claim, tentatively introduced as the claim that all consciousness is self-consciousness in the beginning, amounts to the following: All phenomenally conscious states, all experiences, exhibit mineness, i.e., they are, each and every one of them, given to their subjects as theirs. All of my phenomenally conscious states, in other words, are given to me as mine, all of your phenomenally conscious states are given to you as yours, etc.

Billon and Kriegel put things similarly. They write:

Compare your experiences of drinking apple juice and drinking a banana smoothie. These experiences are different in many respects: there is a gustatory apple-ish way it is like for you to have the former and a gustatory banana-ish way it is like for you to have the latter; there is a tactile juiceish way it is like for you to have the former and a tactile smooth-ish way it is like for you to have the latter, and so on. But there is also one respect in which the two experiences are exactly the same: in both cases it is for you that it is like something to have them. By this we mean not only that both experiences are yours, but more strongly that both are experienced as yours. (Billon & Kriegel 2015, p. 29)

It seems as though no commonly accepted terminology for describing this feature of experiences, i.e., their being given to me as mine, has emerged in the universality debate yet. While Zahavi, in the above quotation, uses the term “mineness”, he also employs different terms, in particular the term “for-me-ness”, in other places. The same goes for Kriegel (see, e.g., Kriegel 2009; Zahavi & Kriegel 2016). By contrast, following Guillot (2017), several other authors (e.g., Farrell & McClelland 2017; O’Conaill 2019; López-Silva 2019) distinguish these two terms, and they reserve the former for our experiences’ being given to us as ours. For the sake of simplicity, I will adopt this stipulation in what follows and refer to my experiences’ being given to me as mine as my experiences’ mineness exclusively. The term ‘for-me-ness’ will be used only when the work of others is referenced.Footnote 7

3 The Transcendental Argument

Now, after this introduction of the universality claim, we come to the vital question of what the arguments are in favor of it: Why should we be universalists? With regard to this question, it must be emphasized that the universality claim is typically framed as a phenomenological claim, i.e., as a claim based on the first-personal investigation of consciousness. Establishing its truth or falsity, hence, is taken to be a matter of getting the phenomenology right, of correctly describing experience from the first-person perspective. As Howell and Thompson put it: “Most of the arguments for the existence of phenomenal meness amount to exhortations to introspection and phenomenological reflection.” (Howell & Thompson 2017, p. 111) And as Farrell and McClelland write with regard to Zahavi’s universalism: “[…] Zahavi suggests that accurate phenomenological description is ‘the best argument to be found’ for thinking that nonreflective experience is characterised by ‘pre-reflective self-consciousness’ of that very experience (2006Footnote 8, p. 24).” (Farrell & McClelland 2017, p. 10; see also López-Silva 2019, p. 323)

As a result of this framing of the debate, much of it has focused on putative phenomenological counterexamples to universalism, i.e., on experiences whose phenomenology does not seem to conform to the universality claim. The most obvious example for such experiences are inserted thoughts in patients suffering from schizophrenia. Hence, it comes as no surprise that the discussion of thought insertion takes up much of the space in the universality debate (see, e.g., Billon & Kriegel 2015; Zahavi 2018; Henriksen et al. 2019). There are, however, other putative phenomenological counterexamples to universalism as well, most notably some of the experiences undergone in meditation or after the ingestion of hallucinogenic drugs (see, e.g., Millière et al. 2018; Letheby 2021; forthcoming).

I will not discuss these debates in what follows. That is, I will not examine the prospects of the universality claim when taken as a phenomenological claim. Neither phenomenological arguments for universalism nor the putative phenomenological counterexamples to universalism just mentioned, i.e., inserted thoughts, etc., will be considered. All of this would simply go beyond the scope of this paper, and there already is an extensive and complex debate on these questions. In view of this, I will focus my attention elsewhere and instead ask what the prospects of universalism are when taken as a non-phenomenological, theoretical claim. More specifically, I will focus on one particular non-phenomenological, theoretical argument for universalism which Zahavi takes to be of crucial importance and which, so far, has hardly been discussed in the universality debate.

Zahavi repeatedly emphasizes that the universality claim is not just a phenomenological claim. In fact, according to him, it is not even primarily a phenomenological claim. As Zahavi puts it: “[…] one should […] not pretend that the issue [of universalism] can be settled simply by […] more detailed experiential descriptions.” (Zahavi 2018, p. 717) And as he also says:

When arguing for the interdependence of consciousness, self-consciousness and selfhood, when arguing that phenomenal consciousness is characterized by for-me-ness and that this amounts to a minimal notion of self, I am not primarily making a descriptive claim such that people who disagreed with me could then be accused of having failed to attend sufficiently carefully to their own experiential life. (Zahavi 2018, p. 717)

Thus, according to Zahavi, the question of universalism cannot be settled by “introspection and phenomenological reflection” (Howell & Thompson 2017, p. 111), “accurate phenomenological description” (Farrell & McClelland 2017, p. 10), etc.: Phenomenology is inconclusive when it comes to universalism. First-personal investigation of experience cannot establish the truth or falsity of the universality claim.

Zahavi also emphasizes that he did not present his own view when claiming that “a correct phenomenological description of our conscious life” is “the best argument to be found” (Zahavi 2005, p. 24) for universalism (Zahavi 2018, p. 711). Instead, he simply outlined a view common among phenomenologistsFootnote 9: Phenomenologists, according to Zahavi, do indeed tend to conceive of the universality claim as forced on us by first-personal investigation, by the attempt to correctly describe experience from the subjective viewpoint. Zahavi, however, goes on to say: “[…] although I have quite a lot of sympathy for that kind of answer, I think a more theoretical argument is needed […].” (Zahavi 2018, p. 711)

The importance Zahavi places on non-phenomenological, theoretical considerations when it comes to defending the universality claim also comes out in the following quotation:

In making these claims [i.e., in arguing for the interdependence of consciousness, self-consciousness and selfhood; see indented quotation above], I have drawn on discussions found not only in contemporary analytic philosophy of mind, but also in Kant, German Idealism, the Brentano school, neo-Kant-ianism, phenomenology, analytic philosophy of language and the Heidelberg School. If one wants to engage in a proper theoretical discussion of for-me-ness, one should engage with these more overarching theoretical discussions […]. (Zahavi 2018, p. 717)

And looking back on an early monograph, Zahavi writes:

In Self-awareness and Alterity, I defended the view that our experiential life is characterized by a form of self-consciousness that is more primitive and more fundamental than the reflective form of self-consciousness that one, for instance, finds exemplified in introspection. In arguing for this claim, I drew on ideas from analytic philosophy of language, post-Kantian German philosophy and phenomenology. (Zahavi 2018, p. 703)

Now, what are the “more overarching theoretical discussions” Zahavi (2018, p. 717) refers to? Which “ideas from analytic philosophy of language, post-Kantian German philosophy and phenomenology” (Zahavi 2018, p. 703) does he draw on? And more specifically, what is the “more theoretical argument” (Zahavi 2018, p. 711) that Zahavi mentions and that he takes to be essential in order to defend the universality claim?

In what follows, I will try to answer these questions. In particular, I will focus on Zahavi’s “more theoretical argument” (Zahavi 2018, p. 711), which he shortly thereafter calls the “central argument” (Zahavi 2018, p. 711) for universalism: I will unfold the argument’s structure, discuss its implications, and examine its scope and strength. My conclusion will be that, while the argument is of considerable force and deserves more attention, it does not lead to universalism in any straightforward way. The argument cannot vindicate the universality claim.

Zahavi introduces the argument in question as “a kind of transcendental argument” (Zahavi 2018, p. 711). And while he does not lay out the exact meaning of the term “transcendental” in this context, the reason why the argument may be said to be transcendental appears quite clear: The argument asks for the conditions of the possibility of, in the version we will discuss, self-identification. And this question for the conditions of possibility seems to make it a transcendental argument.

Zahavi locates the argument’s origins in the philosophical schools and traditions mentioned above, namely “analytic philosophy of language”, “post-Kantian German philosophy”, and “phenomenology” (Zahavi 2018, p. 711). And as notable proponents of the argument in the analytic camp, he mentions “Wittgenstein, Castañeda, Shoemaker” (Zahavi 2018, p. 711), whereas the main continental proponents Zahavi refers to are “Fichte, Henrich and Frank” (Zahavi 2018, p. 711). Of course, the versions of the argument put forward by these authors differ. And in some cases, their differences appear to be quite profound. Zahavi, however, emphasizes the unity behind these differences: According to him, the seemingly unequal group of philosophers mentioned shares the same intuition. Despite their apparent heterogeneity, they stand on a common ground.

Now, in my discussion of the transcendental argument, I will not examine and compare the various different versions of it. I will rather focus on one classical version: the version put forward by Shoemaker. This version seems more accessible than some of the others, it is widely known in contemporary philosophy of mind, and it is repeatedly cited by Zahavi at length. In Subjectivity and Selfhood, e.g., he summarizes Shoemaker’s version of the transcendental argument as follows:

In order to identify something as oneself one has to hold something true of it that one already knows to be true of oneself. In some cases, this self-knowledge might be grounded on some further identification, but the supposition that every item of self-knowledge rests on identification leads to an infinite regress (Shoemaker 1968, 561). (Zahavi 2005, p. 28; see also 2018, p. 716)

The full version of the argument, in Shoemaker’s seminal paper “Self-Reference and Self-Awareness” just quoted by Zahavi, goes as follows:

It is clear, to begin with, that not every self-ascription could be grounded on an identification of a presented object as oneself. Identifying something as oneself would have to involve either (a) finding something to be true of it that one independently knows to be true of oneself, i.e., something that identifies it as oneself, or (b) finding that it stands to oneself in some relationship (e.g., being in the same place as) in which only oneself could stand to one. In either case it would involve possessing self-knowledge – the knowledge that one has a certain identifying feature, or the knowledge that one stands in a certain relationship to the presented object – which could not itself be grounded on the identification in question. This self-knowledge might in some cases be grounded on some other identification, but the supposition that every item of self-knowledge rests on an identification leads to a vicious infinite regress. (Shoemaker 1968, p. 561)

I will use this version of the transcendental argument as the basis of my discussion. Before going into the specifics of the argument, however, the first thing to be noticed is the following: In the passage quoted above, Shoemaker does not mention the concept of phenomenal consciousness at all. The notion of phenomenality does not seem to play any role.

This is remarkable, since universalism, as we have seen, is typically introduced as a claim about phenomenal consciousness: The debate on mineness emerged out of the debate on experience. As the latter debate progressed, a growing number of philosophers, in particular Levine (2001) and Kriegel (2009), began to emphasize that phenomenal consciousness does not just involve qualitative character, i.e., the particular feel that makes a phenomenally conscious state the phenomenally conscious state it is, e.g., the state’s bluishness, its reddishness, etc. In addition to qualitative character, experience also entails subjective character, i.e., the experience’s mineness. Kriegel, thus, writes:

When I have a conscious experience of the blue sky, there is something it is like for me to have the experience. In particular, there is a bluish way it is like for me to have it. This “bluish way it is like for me” constitutes the phenomenal character of my experience. […] [It] has two distinguishable components: (i) the bluish component and (ii) the for-me component. I call the former qualitative character and the latter subjective character. (Kriegel 2009, p. 1)

Claims such as this eventually led to the universality debate, with its forays into psychopathology, hallucinogenics, etc.

In view of this genesis of the universality debate, it is striking that the argument Zahavi highlights as “central” for his universalist view makes no mention of the concept of phenomenal consciousness. Given the close proximity between mineness and experience assumed at the outset of the debate on universalism, the absence of any talk about phenomenality in Shoemaker’s argument is conspicuous.

In fact, it seems as though it is not just Shoemaker’s version of the transcendental argument that is free of any reference to the concept of experience. The same goes, as far as I can see, for the versions put forward by the other philosophers named by Zahavi: None of them seem to ascribe any essential role to the notion of phenomenal consciousness in their versions of the transcendental argument. In fact, this notion does not seem to show up in their versions at all. The named philosophers do, of course, speak of ‘consciousness’, of ‘conscious states’, etc. But they do not specify whether the consciousness they have in mind is of the phenomenal sort, whether they accept the distinction between phenomenal and non-phenomenal consciousness, whether they endorse Nagel’s characterization of phenomenally conscious states in terms of what-it’s-like-ness, etc.Footnote 10

Howell and Thompson take this to constitute a fundamental objection against the use of Shoemaker-style arguments in defense of universalism. Referencing several such arguments, they state that “[p]lenty of accounts of self-consciousness do not give a central role – at least in any obvious way – to phenomenal consciousness” (Howell & Thompson 2017, p. 115). And this, evidently, makes the use of Shoemaker-style arguments in defense of universalism appear problematic. For if universalism aims at establishing an essential connection between phenomenal consciousness and mineness, then it is unclear how arguments in which the concept of phenomenal consciousness is missing could help.

This is a valid point. I believe, however, that the transcendental argument still warrants attention. Despite Howell and Thompson’s objection, we should engage with it in some detail. For if the existence of an essential connection between consciousness and self-consciousness could be shown independently of considerations on phenomenality, what-it’s-like-ness, the possibly unique character of experience, etc., this would still be a highly relevant outcome. If it could be demonstrated that all conscious states, whether phenomenal or not, display mineness, this would certainly be a striking result. In view of this, I will leave the objection that the transcendental argument makes no reference to experience and, therefore, cannot establish an essential connection between phenomenal consciousness and mineness aside in what follows. And I will examine the argument’s strengths and limitations irrespectively of whether or not the concept of consciousness presupposed in it is the phenomenal one.

Also, it should be noted that Shoemaker focuses on a limited number of conscious states, namely those which are self-ascribed by way of first-person statements. And this raises the question of whether his considerations can be generalized: Can they be applied to conscious states which are not self-ascribed first-personally as well? Can they be applied to all conscious states? – Zahavi seems to affirm these questions: He seems to assume that Shoemaker’s version of the transcendental argument has consequences that go well beyond the problem of self-ascription by way of first-person statements. And the reason for this is that, as we will see, the argument reveals a model which, according to Zahavi, cannot only be found in reductive explanations of the first-personal self-ascription of conscious states. Instead, this model, in one version or another, is, supposedly, presupposed in many of the main attempts at explaining conscious states in general reductively, independently of whether or not these states are self-ascribed first-personally.Footnote 11 Therefore, if the model revealed by Shoemaker can be shown to fail, by way of his version of the transcendental argument, then this is a result which has much broader implications: Its significance extends well beyond the admittedly limited number of conscious states Shoemaker is focused on. And because of this, Shoemaker’s argument should not be rejected in advance, due to its limited focus on those conscious states which are self-ascribed by way of first-person statements.

With that said, let us turn to the argument itself: Shoemaker’s main premise is that my identification with a perceived object presupposes a prior self-consciousness on my part. In order to identify myself with the perceived object, I, thus, have to already be self-conscious prior to the identificatory act. And the question, then, is how this prior self-consciousness is brought about, how it can be explained: Is it the result of some further identificatory act? According to Shoemaker, in some unusual circumstances, this may indeed be the case: If, e.g., I lived in a world full of mirrors (Shoemaker 1968, p. 561), my identification with the first object perceived might be mediated by my identification with a second perceived object. I might identify myself with the first object because it stands in a relation of identity to a second object with which I have already identified myself. However, my identification with this second object again presupposes a prior self-consciousness on my part. And this cannot go on indefinitely. Therefore, if we want to avoid an infinite regress, we have to, at some point, assume a self-consciousness that is not grounded on a perception of and identification with any object. We have to, in other words, assume a non-perceptual and non-identificatory self-consciousness as the basis for all of my acts of identification with objects I perceive. Self-consciousness, hence, cannot be construed on the model of perception. The so-called perceptual model of self-consciousness presupposes what it wants to explain, it is circular, and, therefore, fails.

Shoemaker, thus, emphasizes that the structure of self-consciousness is fundamentally different from that of perception: Self-consciousness cannot be reduced to or explained in terms of self-perception. For self-perception involves an “epistemic divide” between the perceiver and that which is perceived, as Zahavi (2005, p. 64) puts it: Self-perception, taken as such, does not bring about the consciousness of the perceiver that that which is perceived is herself, a state of herself, etc. Scenarios such as Mach’s supposed encounter with a shabby schoolmaster who is, in fact, himself, reflected in a mirror, make this abundantly clear (Mach 1886, p. 3): In this scenario, Mach perceives himself without realizing it. And in order to come to the realization that the person he perceives is, in fact, himself, Mach has to have a self-consciousness in the first place, which, at its core, must not leave room for an “epistemic divide”. Thus, self-consciousness, at its core, cannot be structured like self-perception.

Now, Zahavi does not lay out in detail how exactly Shoemaker’s critique of the perceptual model should be applied to the phenomenon of mineness. However, a natural suggestion seems to be this: According to the perceptual model, the mineness of conscious states is supposedly explained as the result of my inwardly perceiving these states and recognizing them as mine. And Shoemaker’s critique of this explanation seems to be that in order to come to this recognition, I have to have some other conscious states that are already given to me as mine in the first place, such as my inner perception of the conscious states in question, etc. The explanation of mineness by way of inner perception, thus, has to presuppose mineness, and, therefore, it fails.

3.1 The Potential of the Transcendental Argument

In my critique in what follows, I will go against Zahavi’s use of Shoemaker’s argument in making his case for universalism. I will contend that the argument does not warrant the universalist conclusions Zahavi draws from it. My critique is not directed against the argument itself, however. I do not reject the argument as such. In fact, I believe that Shoemaker’s argument has considerable force and that its full anti-reductionist potential is often unjustifiably overlooked in the contemporary debate. Thus, while I doubt that the argument’s implications reach as far as Zahavi thinks, I do believe that they reach further than most contemporary philosophers are willing to assume.

Let me quickly hint at what I mean by this: There is a tendency in the contemporary debate to use the terms ‘self-consciousness’ and ‘self-representation’ almost synonymously, to simply presuppose that self-consciousness must be some sort of self-representation (see, e.g., the titles of Newen & Vogeley 2003; Vosgerau 2009). This tendency, however, appears problematic, in view of Shoemaker’s argument. For the main intuition of the argument can, it seems, be applied not just to explanations of self-consciousness in terms of self-perception, but, more generally, to explanations of self-consciousness in terms of self-representation. The argument’s scope, thus, does not appear to be restricted to perceptual models of self-consciousness. Instead, it seems to call into question representational models of self-consciousness in general.

Put simply, the Shoemakerian objection to such models goes as follows: In order to represent an entity as oneself, as belonging to oneself, etc., the representing entity has to have a self-consciousness that cannot, in the end, be explained in terms of self-representation, since no self-representation, taken as such, can guarantee that the representing entity recognizes the represented entity as oneself, as belonging to oneself, etc. Self-representation, thus, is not sufficient for self-consciousness.

Now, I will not discuss this line of reasoning in more detail and apply the transcendental argument to specific explanations of self-consciousness in terms of self-representation. This is not necessary for my overall argument, and I have done so elsewhere (see Wehinger 2016; see also Frank 2012; 2015; 2019; Zahavi 2004). Instead, I will remain focused on the main task of this paper, i.e., the evaluation of the relation between the transcendental argument and universalism, and, more specifically, the question of what the limits of the transcendental argument in the universalist context are.

3.2 The Limits of the Transcendental Argument

Zahavi seems to think that if the explanation of self-consciousness by way of the perceptual model or some version of it fails, then self-consciousness must be taken to be a built-in or intrinsic feature of all conscious states: The failure of the perceptual model, in other words, supposedly necessitates universalism. This comes out quite clearly, e.g., in the following passage:

If reflection always presupposes pre-reflective self-awareness, and if we are capable of reflecting on all our intentional acts and mental states, the conclusion seems obvious: consciousness as such must originally imply self-awareness, since it is impossible to acquire it afterward, be it through reflection or through a study of extrinsic types of self-reference […]. (Zahavi 2020, p. 35)

Zahavi does not speak of perception but rather of reflection in this passage. However, the perceptual model and the reflection model basically amount to the same thing: Both models involve an “epistemic divide” (Zahavi 2005, p. 64) between the subject and the object of perception or reflection. And as a result, both models can be criticized by way of the transcendental argument: This argument shows that in order to recognize a perceived or reflected entity as myself, as belonging to myself, etc., I have to have a non-perceptual or pre-reflective self-consciousness in the first place. And what is crucial in the present context is the conclusion Zahavi draws from this in the quotation above: He takes it to be “obvious” that, in view of the failure of the perceptual or reflection model, “consciousness as such must originally imply self-awareness”. The failure of this model, in other words, supposedly implies universalism.

This line of reasoning will be the main target of my criticism in what follows. I will outline three objections to it. According to the first objection, Shoemaker’s critique of the perceptual model of self-consciousness does not, as such, rule out the possibility of conscious states whose self-ascription fails altogether. And this makes it seem as though Shoemaker’s critique is compatible with the existence of consciousness without mineness. According to the second objection, the applicability of the transcendental argument to certain reductive accounts of self-consciousness appears at least questionable. In particular, the argument does not seem to be geared towards the dispositional accounts of mineness put forward, e.g., by Schear and Carruthers. And according to the third objection, finally, even if antireductionism about mineness could be fully established, universalism still would not follow. In fact, I point out that several proponents of the transcendental argument Zahavi mentions explicitly deny the universality claim.

3.2.1 The Transcendental Argument and the Possibility of Conscious States Whose Self-Ascription Fails

Shoemaker’s version of the transcendental argument is based on his famous observation that certain statements are “immune to error due to a misrecognition of the person” or “immune to error through misidentification relative to the first-person pronouns” (Shoemaker 1968, p. 556).Footnote 12 And the statements he has in mind are, as we have said above, the ones in which we self-ascribe conscious states by way of first-person statements: With regard to such states, according to Shoemaker, there is no room for a misidentification of the subject to whom they belong. I cannot be wrong about who has my conscious states. The examples Shoemaker gives for error-immune statements relative to ‘I’ are ‘I feel pain’, ‘I see a canary’, ‘I am waving my arm’, etc., as opposed to ‘I am bleeding’, ‘I have a beard’, ‘My arm is moving’, etc.: These latter statements lack immunity to error due to a misrecognition of the person. They are not error-immune in the relevant way.

Now, universalism does seem to offer an easy explanation of the immunity to error through misidentification. The universality claim provides a possible basis for this feature of conscious states. For if my conscious states are essentially given to me as mine, the impossibility to misidentify their subject apparently follows. The idea of universal mineness seems to offer a natural foundation for the immunity to error regarding ‘I’.

In view of this, it may appear as though there is a close relationship between Shoemaker’s claim that the self-ascription of all conscious states is error-immune in the way described and universalism. This claim may be taken to suggest the mineness of all conscious states. My main point in what follows will be, however, that even if an affinity between Shoemaker and the universalists is assumed, even if some sort of semblance is granted, Shoemaker’s version of the transcendental argument does not, as such, lead to the universality claim. For while Shoemaker does hold that the self-ascription of every conscious state must be error-immune with regard to ‘I’, this is not the immediate conclusion he draws from his critique of the perceptual model of self-consciousness. His immediate conclusion is, in fact, considerably weaker: In the quotation above, laying out his critique, he writes that, given the failure of the perceptual model of self-consciousness, “not every self-ascription could be grounded on an identification of a presented object as oneself” (Shoemaker 1968, p. 561, my emphasis). Shoemaker, in other words, holds that, in view of this failure, there must be some conscious states that are given to me in a way unlike perception and that are self-ascribed without the interposition of any identificatory act. His immediate conclusion from his critique of the perceptual model of self-consciousness, thus, is that there is a non-perceptual core of self-consciousness, constituted by conscious states whose self-ascription is identification-free.

This, however, makes the attempt to use Shoemaker’s critique in defense of the universality claim appear problematic. For the idea that there is a core of self-consciousness that cannot be reduced to self-perception or self-identification is fully compatible with the assumption that some layers around this core may well be structured quite differently. The conscious states constituting these layers may, indeed, be akin to the physical states we perceive, e.g., when looking in the mirror, in that these conscious states may leave room for errors in identifying the subject to whom they belong. In fact, according to the non-universalist interpretation of thought insertion, this is just what happens in cases of inserted thoughts: Schizophrenic patients suffering from thought insertion have countless experiences which do exhibit mineness. They take most of their thoughts to be not inserted but given to them as theirs. In this sense, much of the patients’ conscious life seems to correspond to the conscious life of persons not afflicted by schizophrenia. It appears to be organized in a similar way. For the conscious states at the core of the patients’ self-consciousness, just like the conscious states of healthy individuals, seemingly, display mineness, they are given to the patients in a non-perceptual manner, and their self-ascription is identification-free. According to the non-universalist reading of inserted thoughts, however, some thoughts in the vicinity of this core are constituted in a different manner: These thoughts are conscious but not given to their subjects as theirs. They lack mineness. And as a result, their self-ascription fails: The patients do not ascribe the thoughts in question to themselves but to someone else, i.e., they take them to be inserted into their mind by another person.

I do not want to endorse or defend the non-universalist interpretation of thought insertion in this context: The relation between thought insertion and universalism is highly complex, and it deserves a much more thorough discussion than I am able to provide here. My point in referring to the non-universalist interpretation of thought insertion is the following: This interpretation seems perfectly compatible with Shoemaker’s critique of the perceptual model of self-consciousness, taken as such. For nothing in Shoemaker’s critique appears to rule out the existence of conscious states without mineness altogether. Instead, as we have seen, it merely leads to the conclusion that some conscious states must be given to us non-perceptually and identification-free. And, possibly, the error-immune self-ascription of these states is based on the mineness they display. Shoemaker’s version of the transcendental argument does not, however, entail the universalist view that mineness is a feature of all experiences. It does not imply that all conscious states must be given to their subjects as theirs. There is, in other words, no direct path from the transcendental argument to the universality claim.

The reason for Zahavi’s assumption that Shoemaker’s argument forces universalism on us seems to be that universalism provides a ‘regress stopper’: As we have seen, the perceptual model of self-consciousness is threatened by an infinite regress. And if self-consciousness is an essential feature of experience, if every experience is accompanied by self-consciousness, then this threat is avoided. The regress is stopped. As Zahavi puts it: “[The self-awareness accompanying every experience] is not itself a separate experience in need of yet another awareness. The self-awareness of the experience is an intrinsic, non-reflective, irrelational feature of the experience itself, and thus the regress is stopped.” (Zahavi 2020, p. 36) As we have just pointed out, however, we do not need universalism in order to stop the regress: Universalism is not the only option. In fact, the assumption of a non-perceptual core of self-consciousness, i.e., the assumption that some conscious states are given to me as mine, in a non-perceptual, pre-reflective manner, seems to be enough: This assumption seems to be all that is needed in order to avoid the threat of an infinite regress.

3.2.2 The Transcendental Argument and the Dispositional Accounts of Mineness

At the beginning of Sect. 3, we have seen that Zahavi seems to assume the following conditional: If we cannot explain reductively how conscious states acquire mineness, then mineness must be taken to be an intrinsic feature of experience. And this seems to amount to the claim that if reductionism about mineness can be shown to fail, then universalism follows: If mineness cannot be reduced, then universalism must be true. In the next section, i.e., in Sect. 3.2.3, I will criticize this conditional. Before that, in this section, however, I want to point to the fact that, even if we accept Zahavi’s conditional, the transcendental argument is unable to establish its antecedent. That is, the transcendental argument is unable to establish the failure of reductionism.

And the reason for this seems quite obvious: While the transcendental argument can be applied rather straightforwardly to certain reductive accounts of mineness, as outlined above, its applicability to other such accounts seems much more doubtful. The argument does offer a forceful objection against some attempts to reduce my conscious states’ being given to me as mine. It appears to be an open question, however, whether it can be used against all.

Let me illustrate this by way of the dispositional accounts of mineness espoused by Schear (2009) and Carruthers (2000): According to these accounts, my conscious states’ being actually given to me as mine is explained in terms of their disposition to be so given. Self-consciousness, hence, is conceived as a dispositional feature, “a potentiality – generally unactualized, but always actualizable” (Schear 2009, p. 99). Or as Zahavi and Kriegel put it in their depiction of the dispositional accounts: “Every experience includes a potential for the experience to be for me, not actual for-me-ness.” (Zahavi & Kriegel 2016, p. 47).

Now, my goal is not to assess this position. I merely want to point out that the applicability of the transcendental argument to dispositional accounts of mineness appears at least questionable. The argument simply does not seem to be tailored towards them. For both the explanandum and the explanans in the theories of self-consciousness criticized by Shoemaker are taken to be actually existing: Shoemaker’s argument is directed against theories in which actual self-consciousness is explained in terms of actual self-identification. It can be applied in a rather straightforward way to accounts which reduce actual mineness to actual self-representation, etc., as we have seen. Once dispositional features are introduced as explanans, however, the dialectics of the debate seems to change. The applicability of the transcendental argument to the resulting accounts is no longer evident.

More specifically, there is at least no obvious threat of infinite regress or circularity in the way outlined by Shoemaker when actual mineness is explained in terms of my experiences’ tendency to be given to me as mine, i.e., in dispositional terms. The objections that Shoemaker makes to explanations of self-consciousness in terms of self-perception do not seem to be transferable to dispositional accounts of mineness directly. Of course, several other objections might be made: It might be said, e.g., that dispositional explanations in general are vacuous, that they fall prey to the virtus dormitiva-objection, and that they must be replaced by categorical explanations, since it is categorical properties that do all the causal work in the world. What must be emphasized, however, is that these objections are quite distinct from the ones Shoemaker raises. They are not based on the peculiarities of first-person reference, on the non-perceptual character of self-consciousness, etc. Instead, they are premised on claims concerning the nature of dispositions, their relation to categorical properties, etc. When raising these objections, we are, thus, leaving the realm of philosophy of mind and entering the realm of metaphysics and philosophy of causation.

The impression that the transcendental argument is not suited for a rebuttal of dispositional accounts of mineness gets further confirmation from the fact that Zahavi and Kriegel do not resort to this argument but rather to metaphysical considerations concerning the dispositional/categorical-distinction quite similar to the objections just mentioned in their response to the dispositional accounts: Their response is based on the metaphysical claim that dispositional features, ultimately, need categorical bases. The categorical, in other words, has ontological priority over the dispositional. As Zahavi and Kriegel put it: “Plausibly, […] dispositional phenomena always presuppose categorical bases, so in the vicinity of every dispositional explanandum there must also be a categorical explanandum that underlies it.” (Zahavi & Kriegel 2016, p. 47)

Now, the categorical base that Zahavi and Kriegel propose in the case of the disposition for self-consciousness, unsurprisingly, is “the ubiquitous for-me-ness of experience” (Zahavi & Kriegel 2016, p. 47). The latter feature, according to them, grounds the former, and not vice versa, contrary to what is claimed in dispositional accounts. The reason why Zahavi and Kriegel take their position to be explanatorily superior to that of their opponents, hence, is that it can satisfy the metaphysical principle that dispositions must, ultimately, be explained in categorical terms. As they put it: “There is here an undeniable gain in explanatory depth, since in general the dispositional can be explained in terms of the categorical but not the other way round […].” (Zahavi & Kriegel 2016, p. 48).

It is not my goal to discuss the validity of this proposal. What I want to point out is simply that Zahavi and Kriegel’s objection to the dispositional accounts of mineness is in no way based on the transcendental argument. And this seems to confirm the impression that the argument is not of much help against these accounts.

Zahavi, thus, appears to be aware of the limitations of the transcendental argument, when taken as an argument against reductionism about mineness. Some of his formulationsFootnote 13, however, as well as his emphasis on the centrality of the transcendental argument may be taken to suggest that this argument is all that is needed in order to establish antireductionism. And what I wanted to point out in this section is that this is not the case.

3.2.3 Antireductionism About Mineness Does Not Imply Universalism

While the second objection above is aimed at the antecedent of Zahavi’s conditional, i.e., the claim that if reductionism about mineness can be shown to fail, then universalism follows, the third and probably most fundamental objection I want to make calls into question this conditional as a whole: My point is that even if all reductive accounts of self-consciousness could be refuted, by way of the transcendental argument or some other objection, even if it could be shown that mineness cannot be reduced in any way, universalism still would not follow. Even if antireductionism about self-consciousness could be established without qualification, this would still not be enough to vindicate the universality claim.

In the passage quoted at the beginning of Sect. 3, Zahavi claims that “consciousness as such must originally imply self-awareness, since it is impossible to acquire it afterward, be it through reflection or through a study of extrinsic types of self-reference” (Zahavi 2020, p. 35). Zahavi, thus, seems to assume that either self-consciousness can be explained reductively or it must be taken to be an intrinsic feature of experience: Either reductionism can be established or universalism must be true.

This appears to be a non sequitur, however: Antireductionism about self-consciousness simply does not imply universalism. Non-universalist antireductionism is a perfectly consistent position. There is no necessary connection between antireductionism and universalism. One can be an antireductionist about self-consciousness without claiming that all consciousness is self-consciousness.

Zahavi seems to assume that the alleged irreducibility of mineness forces the universality claim on us, that once this irreducibility is established, universalism is the only viable path. But this assumption appears highly questionable. In fact, a closer look at the accounts of the authors Zahavi mentions as the main proponents of the transcendental argument seemingly confirms the impression that there is no necessary connection between antireductionism about mineness and universalism: All of these authors have antireductionist tendencies or fully embrace antireductionism about self-consciousness.Footnote 14 Nevertheless, several of them explicitly reject the universality claim. They deny the mineness of all experiences in no uncertain terms.

Castañeda, e.g., writes: “[…] contrary to a widely held Fichtean tradition, not all consciousness is self-consciousness. […] self-consciousness is built upon layers of self-less consciousness […].” (Castañeda 1989, pp. 29–30) And he goes on: “[…] there are episodes of I-less consciousness. It is not just that there is a quiet I in the background, which can at any moment come into the open.” (Castañeda 1989, p. 51)

Henrich is equally adamant in his denial of the universality claim. He writes, e.g.:

If one asks how […] consciousness is to be understood, it is no longer “obvious” that it belongs to an ego, and hence is basically self-consciousness. We know, to be sure, that it is a person who begins to awaken or to dream, but this does not mean that the consciousness-structure which we attribute to an animal organism or a person capable of speech and action must necessarily be related to a self. Indeed, there is nothing in the facts as we know them to support this claim. (Henrich 1971, p. 6)

Similarly, Frank holds that “the I is no native resident of the anonymous field of consciousness” (Frank 2012, p. 357, my translation). And he continues:

Instead of saying ‘I’m in pain’, we should, at this level, say ‘It hurts’. A good example for the anonymity of the field of consciousness prior to the intervention of attentional and cognitive acts is the awakening from an anesthesia: We feel exposed to a tangled web of dull bodily sensations, pains, smells, colors, perhaps distorted faces, and we do not, ourselves, appear in this picture. ‘It’ is uncanny, ‘it’ feels strange. Only when we perform a kind of Kantian synthesis of apperception does an I come into play, and we can slowly form objective judgments. (Frank 2012, p. 357, my translation; see also p. 335)Footnote 15

Hence, all of the cited authors, despite their endorsement of the transcendental argument and the irreducibility of mineness, assume that there are some self-less experiences. These experiences are anonymous, they are not given to any selves as theirs. Consciousness, according to Castañeda, Henrich, and Frank, thus, does not entail self-consciousness. The irreducibility of mineness does not imply its universality. Mineness, in this view, is an irreducible feature that some experiences have. But it is not a feature of all experiences. The cited authors, hence, are proponents of the transcendental argument. But they are not advocates of universalism. Their antireductionism about mineness does not lead them to adopt the universality claim.

At this point, it might be interjected that even if there is no necessary connection between antireductionism about mineness and universalism, antireductionism at least suggests the universality claim: Even if non-universalist antireductionism is a logically and metaphysically consistent position, universalist antireductionism is still to be preferred. This claim may be supported by considerations along the following lines: In non-universalist versions of antireductionism, mineness, ultimately, remains ungrounded. It ‘pops into existence’ somewhat unexpectedly, as an addition to phenomenal consciousness. As a result, a gap opens up between experience and self-consciousness. And it seems as though this gap cannot be closed. In universalist versions of antireductionism, by contrast, there is no gap separating phenomenal consciousness and mineness. Since mineness is an essential feature of consciousness, it emerges in conjunction with experience. And as a result, there is a continuum between the most basic kinds of experience and the subtleties of first-person reference. Mineness is not introduced as a completely new feature that only some experiences have but as an essential component of phenomenal consciousness.

This alleged advantage of universalist antireductionism over non-universalist antireductionism, however, comes at an obvious price. For it forces us to embrace the claim that even the most rudimentary and outlandish experiences have mineness, that even phenomenally conscious states which are conceivably far removed from the conscious states of healthy adult humans are given to their subjects as theirs. And it appears questionable whether the benefit of having a continuous view of the mind can outweigh the drawback of being committed to this claim.

Also, evidently, universalist antireductionism is gappy as well: Although it denies a gap between consciousness and self-consciousness, as an antireductionist position, it is committed to the claim that mineness cannot be explained without any gaps. The only difference to non-universalist antireductionism seems to be that this gap is located in a different place: According to universalist antireductionists, the gap separating mineness from the rest of nature opens up with the emergence of experience, since experience and mineness go hand in hand. According to non-universalist antireductionists, by contrast, experience and mineness are separated by a gap. In view of this, it does not seem to be a promising strategy for universalist antireductionists to accuse their non-universalist counterparts of gappyness. For this accusation, obviously, can be made against themselves as well.

4 Conclusion: The Transcendental Argument Cannot Tip the Scales

Summing up, the value of the transcendental argument for universalism appears to be limited. The argument speaks against some, but not all reductive explanations of self-consciousness. And it neither rules out the possibility of inserted thoughts without mineness nor the possibility of self-less experiences in general. In the end, antireductionism about mineness simply does not imply the universality claim.

Now, Zahavi seems to concede that the phenomenological investigation of consciousness is inconclusive with regard to universalism. In fact, this appears to be the reason why he places such importance on the transcendental argument, why he takes it to be so “central” for the universality claim: The transcendental argument, seemingly, is supposed to break the phenomenological stalemate between universalism and non-universalism. It is supposed to tip the scales.

As I have argued, however, the transcendental argument cannot satisfy these expectations. The conclusions Zahavi draws from it do not seem to be warranted. And this puts universalism in a difficult position. For if neither phenomenology nor the best non-phenomenological, theoretical argument can vindicate it, then its prospects do not seem good. In the end, given the problems I have pointed out in the argumentative strategy for universalism, the claim that all experiences have mineness does not appear well-justified. In view of the objections and counterarguments, it must be called into question.