On the Conceivability of a Cognitive Phenomenology Zombie

The cognitive phenomenology thesis has it that conscious cognitive states essentially exhibit a phenomenal character. Defenders of ‘conservatism’ about cognitive phenomenology think that the phenomenology of thought is reducible to sensory phenomenology. In contrast, proponents of ‘liberalism’ hold that there is a proprietary, sui generis cognitive phenomenology. Horgan develops a morph-sequence argument to argue for liberalism. The argument is based on the conceivability of a cognitive phenomenology zombie, i.e. a man who does not understand Chinese but shares the behavior and sensory phenomenology with his twin who does understand Chinese. I argue that the conceivability of a cognitive phenomenology-zombie fails to settle the debate between conservatives and liberals. The roots of the ineffectiveness of the argument lie in the diverse readings of sensory phenomenology which flesh out the relation between sensory phenomenology and concepts differently but explain the conceivability of the scenario equally well. The lesson to learn is that to adjudicate the debate about cognitive phenomenology, we first have to clarify the notion of sensory phenomenology.


Introduction
In the last century, most philosophers endorsed separatism between phenomenal states that essentially exhibit a phenomenal character on the one hand, and cognitive states that do not essentially exhibit a phenomenal character on the other. This separatist picture is challenged by the cognitive phenomenology thesis (hereinafter: CP-thesis). The central claim of the CP-thesis is that occurrent conscious cognitive statesin short, thoughts (understood as tokens of conscious cognitive states)also essentially exhibit a phenomenal character. The CP-thesis comes in different versions and strengths.
One way to distinguish CP-theses is by focusing on which kind of phenomenology figures in conscious thoughts. 'Conservatism' 1 about cognitive phenomenology has it that the phenomenal character of conscious thought is sensory phenomenology (i.e. the phenomenology of sensations and their analogs in imagery) (see Prinz 2011;Robinson 2011;Tye and Wright 2011;Pautz 2013;Carruthers and Veillet 2017).
On the alternative view, labelled 'liberalism', the phenomenology of thought is of a proprietary and sui generis kind. (Liberalism allows two readings. On one reading, cognitive phenomenology is seen as irreducible to the familiar kind of sensory phenomenology (e.g. Horgan 2013). I label this view 'liberalism+'. On a stronger reading, cognitive phenomenology is seen as modally independent of sensory phenomenology (e.g. Pitt 2004;Kriegel 2015). I label this view as 'liberalism*'. 2 These specifications of liberalism will become relevant in Section 3.) A second way to distinguish CP-theses is by focusing on the role the phenomenology plays. What I call the 'strong' CP-thesis has it that phenomenology is individuative or constitutive of conscious thought. That means it is due to the phenomenology that a conscious thought is the specific thought it is. For example, Pitt holds that "the phenomenology of a thought constitutes its representational content (i.e., is individuative)" (2004,5). In contrast, defenders of the 'weak' CP-thesis deny that the phenomenal character of conscious thought has any individuative powerthey rather hold, for instance, that the phenomenal character is merely accompanying conscious thought (e.g. Robinson 2005). For example, Tye and Wright emphasize that they "deny [that] what it is like for a subject when she undergoes a thought is proprietary and further distinctive and individuative of that type of thought" (2011,328).
For the purpose of this paper, I will set the weak CP-thesis aside. The reason is that in this paper I investigate the conceivability of a cognitive phenomenology zombiea scenario that aims at establishing a phenomenology that is individuative of conscious thoughts. Accordingly, I will focus on the strong CP-thesis. In particular, I am concerned with the dispute between liberal and conservative advocates of the strong CP-thesis; i.e. those philosophers who share the view that phenomenology individuates conscious thoughts, but who disagree about what is the apt characterization of the relevant phenomenology. 3 2 A defender of liberalism* is, for example, Pitt (2011, 141): (T)he phenomenology of occurrent conscious thought is proprietary: it's a sui generis sort of phenomenology (…)a cognitive phenomenology. I believe this because I believe that the conscious occurrence of any of the more familiar sorts of phenomenal properties is neither necessary nor sufficient for the occurrence of conscious thought. 3 At this point we see that the cognitive phenomenology debate challenges the orthodox division between cognition and sense-perception. Aptly characterizing the boundary between cognition and perception is a difficult task. One might focus on what perceptual and cognitive states represent and contrast the low-level contents of sensory experiences with the high-level contents of thoughts. Alternatively, one might oppose perception and cognition by considering non-conceptual content versus conceptual content. I discuss these oppositions in Sections 3 and 4. (For a discussion of these and further approaches, see Kriegel (forthcoming) who thinks that none of these divisions work.) The cognitive phenomenology debate is not the only one challenging a sharp distinction between perception and cognition. Further challenges come from considerations about the cognitive penetrability of perceptual states (see Zeimbekis and Raftopoulos 2015) and from recent work on predictive processing (Hohwy 2013;Clark 2016).
A defender of strong liberalism 4 is, for instance, Pitt who summarizes his view as 'thesis P': P: Each type of conscious thoughteach state of consciously thinking that p, for all thinkable contents phas a proprietary, distinctive, individuative phenomenology. (Pitt 2004, 4) On this view, to individuate conscious thoughts, the relevant phenomenology has to be intimately tied to the (narrow) contents which are consciously entertained. Most defenders of liberalism hold that the relation between cognitive phenomenology and content is a constitutive one.
Defenders of a strong conservatism deny the existence of such a proprietary cognitive phenomenology and hold that rather sensory phenomenologye.g. together with functional facts (Pautz 2013)accounts for the constitution of conscious thought. Conservatives think that what speaks in favor of their view is that it relies on the fairly uncontroversial sensory phenomenology and that it is a parsimonious view about the phenomenology of conscious states (Prinz 2011).
Since the very existence of a proprietary, sui generis, cognitive phenomenology is highly controversial, liberals put forward arguments of various forms to establish it. The key arguments for proprietary cognitive phenomenology rely on the method of phenomenal contrast 5 and offer scenarios of the following structure: first, readers are invited to imagine two mental states that differ in their overall phenomenal character. Second, by pointing out that the same sensory phenomenology is involved in these states, it is argued that the phenomenal contrast can only be accounted for in terms of a proprietary cognitive phenomenology. 6 To establish a strong liberalism, namely that proprietary cognitive phenomenology is individuative, a further argumentative step has to be taken. The argument from phenomenal contrast has to incorporate the additional claim that the phenomenal 4 Liberalism typically comes with a strong CP thesis. Strong liberalism is defended, e.g., by Siewert (1998), Pitt (2004, Strawson (2009), Horgan (2013), Chudnoff (2015, Kriegel (2015), and Montague (2016). However, liberalism does not entail the strong CP thesis and could be combined with the claim that proprietary cognitive phenomenology is not individuative of conscious thought. 5 'Phenomenal contrast arguments' are famously found in Siegel (2010). Examples that employ the method of phenomenal contrast to argue for cognitive phenomenology can be found, for instance, in Siewert (1998Siewert ( , 2011, Horgan and Tienson (2002), Pitt (2004), Strawson (2009), Horgan and Graham (2012), Chudnoff (2015, and Kriegel (2015). 6 This holds under the assumption that no further kind of phenomenology exists that is neither sensory nor cognitive. If there is, e.g., an agentive phenomenology (Horgan 2011) that is neither reducible to sensory nor to cognitive phenomenology, the argument would have to exclude this phenomenal feature too.
On the Conceivability of a Cognitive Phenomenology Zombie 107 contrast is content-specific. If the phenomenal contrast scenarios succeeded in meeting both aims, strong liberalism would be established. The primary target of this paper is the dispute between liberals and conservatives. In particular, I will demonstrate that those arguments from phenomenal contrast that involve sensory phenomenology cannot settle the issue whether sensory phenomenology or proprietary cognitive phenomenology is individuative of conscious thought. I will argue for this claim by analyzing Horgan's (2013) "morph-sequence argument". Horgan asks us to imagine a man who does not understand Chinese but who is in several important respects identical to one who does. The argument involves a series of thought experiments that aim to establish the existence of a proprietary and individuative cognitive phenomenology. Although the discussion in this paper focuses mainly on Horgan's thought experiment, it aims to illuminate the general issue of how liberals and conservatives understand the relation of cognitive phenomenology to sensory phenomenology. Clarifying the notion of sensory phenomenology is a challenge that applies to all those who claim or deny that cognitive phenomenology is reducible to sensory phenomenology.
I proceed as follows: Section 2 summarizes the main points of Horgan's argumentation and characterizes the notion of a "cognitive phenomenology zombie". Section 3 investigates which versions of a cognitive phenomenology zombie are coherently conceivable. I argue for a neutral interpretation of the scenario that does not assume away the problem. Section 4 argues that the neutral interpretation of Horgan's scenario fails to establish liberalism, since a less committal conservatism that explains the conceivability of the scenario equally well is near at hand.
2. An argument from phenomenal contrast: Horgan's cognitive phenomenology zombie In the paper "Original Intentionality Is Phenomenal Intentionality" (2013), Horgan develops a series of scenarios inspired by Searle's (1980) Chinese room experiment to "offer a new argument in favor of cognitive phenomenologyand, more specifically, in favor of the (distinctive, proprietary, individuative) phenomenology of language-understanding" (2013, 241). To reach his target, Horgan embraces an argument from phenomenal contrast. Horgan presents the following scenario: a mancalled "the guy"who does not understand Chinese, undergoes different stages and degrees of differences to a person who does understand Chinese. In particular, each stage provides him with more behavioral and sensory-phenomenal elements that are shared with his Chinese-understanding twin. 108 Martina Fürst The series starts with Searle's Chinese room scenario where a person is manipulating symbols he does not understand. In the next stage, the manipulation of the symbols is done not by the guy himself, but (very rapidly) by a monitoring/processing/stimulation device (MPS device) appended to the guy's brain. The MPS device monitors the visual input coming into the guy's eyes, takes note of the input symbols (in Chinese) the guy sees, rapidly and automatically executes the symbol-manipulation rules, and then stimulates the guy's brain in a way that produces totally spontaneous decisions to put certain (Chinese) symbols into a box. (Horgan 2013, 242-243) Note that, besides having the MPS device implanted in his head, there is a further important difference between the guy and his Chinese-speaking twinthe guy suffers from a significant cognitive deficit that explains why he is unable to learn Chinese 7 : the guy also has a serious memory deficit: he persistently lacks any memories (either episodic or declarative) that extend further back in time than thirty seconds prior to the current moment. Because of this, he is unable to learn any Chinese on the basis of what he sees and hears. (Horgan 2013, 242) The stages described by Horgan get progressively more complex, involving more and more factors that the guy shares with a person who does understand Chinese. The fifth and final stage is supposed to demonstrate that the only reason why the guy still lacks any understanding of Chinese is his lack of proprietary cognitive phenomenology. Horgan describes the key scenario in the following way: Stage 5: (1) The MPS device now monitors all the guy's sensory inputs (not just visual or auditory inputs). It also monitors all his occurrent desires and beliefs and other mental states (both present and past). It constantly stimulates his brain in ways that generate spontaneous decisions (or seeming-decisions) on his part to move his body in ways that are suitable to the overall combination of (a) the guy's beliefs and desires and other mental states (both present and past, many of which are, of course, currently forgotten by the guy himself) and (b) the content of his current sensory input (including the content of the meaningless-to-him sign-designs in his sensory input that happen to be written Chinese or spoken Chinese).
(2) The MPS device generates in the guy any (noncognitive) sensory images, (noncognitive) emotional responses, and other noncognitive phenomenology that would arise in a guy who (a) understood Chinese, (b) had normal memory, and (c) was mentally and behaviorally just like our guy.
(3) The MPS device prevents from occurring, in the guy, any conscious mental states that would normally, in an ordinary person, accompany mental states with features 7 As we will see in Section 4, the guy's memory deficit plays a crucial role in an explanation of the conceivability of the scenario.
On the Conceivability of a Cognitive Phenomenology Zombie 109 (1)-(2) (e.g., confusion, puzzlement, curiosity as to what's going on, etc.). This includes precluding any noncognitive phenomenology that might attach to such states.
(4) Rather than being stuck in a room, the guy is out among the Chinese population, interacting with them both verbally and nonverbally. He is perceived by others as being a full-fledged, ordinary, fluent speaker of Chinese. This guy understands no Chinese at all. Each of these successive stages is coherently conceivable, I submit. And for each scenario, it seems obvious that the guy understands no Chinese. (Horgan 2013, 243) The final stage of the scenario contains substantial modifications. For our purposes, the most important modification is that now the guy also shares the "sensory images, emotional response and other non-cognitive phenomenology" with a person who understands Chinese, but still does not understand Chinese.
The structure of the argument is the following. First, Horgan's scenario is supposed to evoke the intuition that there is a phenomenal contrast between the guy and his Chinese-understanding twin. Second, this phenomenal contrast is held to be due to the guy's lack of understanding the target thoughts. From this Horgan concludes that a proprietary phenomenology is involved in entertaining a specific thought. That means, according to Horgan, the best explanation for the intuition that there is a phenomenal contrast between the mental states of the guy and those of the Chinese speaker is that the guy lacks a proprietary cognitive phenomenology. 8 Thus, the morph-sequence argument is an abductive argument that aims to establish a proprietary cognitive phenomenology that is irreducible to sensory phenomenology and that is individuative of conscious thoughts (i.e. a strong liberalism).
Before proceeding, let me draw the attention to a decisive challenge that is faced by Horgan's argument and that motivates the analysis in Section 3. The concern is that for an abductive argument to succeed, first, we have to share the main intuition which, in the following step, is supposed to be explained best by the target thesis. More specific: first, we have to agree that the scenario described by Horgan is conceivable and second, we have to agree that its conceivability is best explained by the existence of a proprietary cognitive phenomenology. In Section 3, I focus on the first step. Since Horgan's scenario is in one important respect under-described, it allows more than one reading. In particular, different interpretations of 'sensory phenomenology' will result in opposing intuitions regarding the conceivability of the scenario. The worry is that without a shared intuition of the phenomenal contrast between the guy's mental states and the native Chinese speaker's mental states, the morph-sequence argument will not get off the ground.

Ways of conceiving of Horgan's CP-zombie scenario
To analyze Horgan's argument, I start by investigating ways of conceiving of the scenario in which "it seems obvious that the guy understands no Chinese" (Horgan 2013, 243). Since the man in Horgan's scenario lacks understanding experiences, 9 he can be seen as a version of a "cognitive phenomenology zombie". What I call hereinafter a "CP-zombie" is a creature that differs in some respects from zombies in the orthodox zombie arguments against physicalism (Block 1980;Chalmers 1996). First, a CP-zombie is not functionally identical, but only behaviorally isomorphic to his Chinese-understanding twin. 10 Second, and importantly, the CP-zombie also shares all sensory phenomenal states with his twin. That means he has experiences of seeing colors, smelling roses, tasting red wine and also experiences of inner images, subvocalization, emotional and bodily responses. The crucial point here is that the CP-zombie supposedly lacks any genuine cognitive phenomenology tied to Chinese. Thus, following Horgan's description of the scenario, I will use the notion "CP-zombie scenario" as abbreviation for the conjunction of the following three claims: the man in the CP-zombie scenario is (1) behaviorally isomorphic to his Chinese speaking twin, (2) he shares all sensory phenomenology with his Chinese speaking twin, and (3) he lacks understanding experiences with respect to Chinese.
The conceivability of the CP-zombie scenario is supposed to achieve two things. First, it should speak in favor of a proprietary cognitive phenomenology. If there is a kind of phenomenology that is conceivably separable from behavioral and sensory-phenomenal properties, it is possible that it is separable from behavioral and sensory-phenomenal properties. One might conclude that cognitive phenomenology is not identical with or supervenient on behavioral and sensoryphenomenal properties. 11 Second, the conceivability of a CP-zombie also aims 9 In the literature, the notion of 'understanding-experiences' is sometimes used to refer to 'aha!'-experiences or a general feeling of getting it. In contrast, Horgan uses the notion of 'understanding-experiences' to point at the phenomenal character of having a particular conscious thought. Accordingly, when we are told that the guy lacks understanding experiences, this means that he is not consciously entertaining the specific thoughts that his Chinese-speaking twin is entertaining.
10 Some further clarifications of the similarity-claim are called for. Obviously, the guy is not physically identical to someone who does understand Chinese, since he operates via the MPS device. (However, I doubt that the conceivability of Horgan's guy crucially hinges on him having the MPS device in his head.) Moreover, since Horgan does not think of the MPS device as part of the guy's mind, the guy also does not have internal states that are functionally similar to those of a Chinese-understanding person. The isomorphism concerns only the guy's behavior, characterized in bodily motion terms. This is a further aspect that differentiates CP-zombies from the orthodox zombies (who are stipulated as functionally identical to their human twins). 11 There is significant controversy about whether conceivability entails possibility. For a defense, see e.g. Chalmers (2010); for criticism of this entailment, see e.g. Yablo (1993), Block and On the Conceivability of a Cognitive Phenomenology Zombie 111 to establish an individuative cognitive phenomenology since what the zombie lacks is not just a general feeling of understanding. Let me explain. Horgan explicitly tells us that we should not conceive of the guy as puzzled, bewildered, or wondering what is going on (Horgan 2013, 243). Accordingly, we should not conceive of the guy as just lacking a general "aha!"-feeling (that might be equally tied to all thought types) or a phenomenology that would be close to emotions such as lacking a feeling of fluency or being puzzled when he hears or produces Chinese utterances. Rather, the CP-zombie is supposed to lack a kind of phenomenology that is content-specific and, hence, individuative of thoughts.
Can we coherently conceive of a behavioral and sensory-phenomenal twin of a Chinese speaker who lacks understanding experiences with respect to Chinese? To answer this question, we have to be clear about the notions of behavioral and sensory-phenomenal similarity that are at issue here.
The CP-zombie is supposed to be behaviorally isomorphic to his Chinese-understanding twin. In contrast to the orthodox zombie-scenario which involves functional identity, assuming a behavioral isomorphism is a much weaker claim. 12 So the key issue is: how should we understand the claim that the CP-zombie shares all sensory phenomenology with his Chinese-understanding twin? In Horgan's argument, as well as in many other arguments from phenomenal contrast, the notion of sensory phenomenology is underspecified. 13 So what notion of sensory phenomenology can be used to conceive of the described scenario?
There are at least two ways of further specifying the notion of sensory phenomenology. One specification concerns the very nature of sensory phenomenology, and distinguishes between conceptual and non-conceptual sensory phenomenology. Another way to flesh out diverse readings of sensory phenomenology is to focus on the relation in which concepts stand to sensory phenomenology. The opposition then turns on the question whether sensory phenomenology is conceptualized in a constitutive or in a causal way. 14 In this section, I focus on the first distinction (viz. conceptual versus non-conceptual sensory phenomenology) in order to motivate a neutral interpretation of the sensory phenomenology that is compatible with both conservatism and liberalism. Once we have found a neutral way of conceiving of the CP-zombie, we can take this as a starting point for investigating whether conservatism and liberalism explain the scenario equally well. Conservatism and liberalism might explain the scenario differently by focusing on different ways in which the sensory phenomenology is conceptualized. Accordingly, in Section 4, I consider the kind of impact that concepts have on the sensory manifold and focus on the second distinction (viz. causal versus constitutive conceptualization) of sensory phenomenology.
The debate about the conceptuality of the content of sensory experiences is lively and sophisticated. 15 To analyze the variety of accounts on the topic (and how they relate the content and the phenomenology of sensory states) in detail would carry us too far off course, though. What matters for the present purpose is that two rival readings of 'sensory phenomenology'those that lie at the opposite ends of the spectrumare problematic if employed in Horgan's argument. Chudnoff (2015) discusses the general difficulty of the competing low-level/high-level interpretations of sensory phenomenology in detail. He thinks that for the notorious difficulty to agree on a shared characterization of sensory phenomenology, phenomenal contrast arguments involving understanding experiences are problematic. Thus, Chudnoff develops an argument for cognitive phenomenology which is based on mathematical intuitions and evades the problem by not relying on sensory phenomenology. Since the main objective of this paper is to elaborate on the limits of arguments from phenomenal contrast that are based on sensory phenomenology, I leave the discussion of Chudnoff's 'glossed phenomenal contrast argument' to another paper. 14 The relation between these two kinds of opposing readings can be fleshed out in different ways. For example, Carruthers and Veillet (2011) think the conceptual/non-conceptual distinction is orthogonal to the causal/constitutive distinction. 15 For a collection of papers on the contents of sensory experiences, see Hawley and Mac-Pherson (2011). Arguments for the non-conceptual content of sensory states encompass the fineness of grain of sensory experiences (Evans 1982), their analog character and informational richness (Dretske 1981) and the fact that they can represent contradictory states of affairs (Crane 1992). Arguments for the conceptual content of experiences are mostly based on the view that perceptual experience states can provide reasons for beliefs (McDowell 1994;Brewer 2005).
On the Conceivability of a Cognitive Phenomenology Zombie 113

Interpretation (A): Conceptual sensory phenomenology
One reaction of conservatives to the scenario described by Horgan is to hold that it is inconceivable. One way to arrive at this result is to construe sensory phenomenology as conceptual. In particular, one might hold that the sensory phenomenology tied to occurrent thoughts is conceptual in a very specific waynamely in such a fine-grained way as to provide the subject with the specific thought-content. Why should we think that sensory phenomenology is conceptual in such a way? Recall that according to conservatism, sensory phenomenology is the only phenomenology there is. Accordingly, as far as phenomenology is concerned, sensory phenomenal similarity is similarity in conscious thoughts simpliciter. So, when asked to conceive of Horgan's CP-zombie scenario, conservatives might react in the following way: if the CP-zombie shares all sensory phenomenal properties with his Chinese-understanding twin, he will also have the conscious target thought. One way to achieve this result is by holding that sensory phenomenology is conceptually structured in such a specific way as to entail the target thought.
Unfortunately, an interpretation of the scenario that stipulates such specific conceptual sensory phenomenology is problematic for several reasons. Firstly, it assumes away the problem at hand. It is true that we cannot coherently conceive of a CPzombie with the sensory phenomenology of a sub-vocalization conceptualized as the specific words with the meaning "There is a bottle of wine in the fridge", the inner image of a bottle of wine in the fridge and the bodily movement to intentionally go and get the bottle of wine in the fridge, but yet without having any clue about the content of the thought. Rather this specifically conceptual sensory phenomenology entails that the guy will have the target thought. However, one assumes away the very problem if one simply holds that (a) Horgan's CP-zombie scenario is inconceivable and (b) the reason for its inconceivability is that sensory phenomenology is conceptually structured in a specific way to entail the target thought. Moreover, this interpretation fails to explain the intuition of many philosophers that they can conceive of the CP-zombie lacking understanding experiences. 16 Secondly, it is far from clear whether conservatives should analyze cognitive phenomenology in terms of such specific conceptual sensory phenomenology. It 16 The orthodox zombie argument against physicalism elicited different reactions. Some functionalists simply deny the conceivability of zombies, turning the situation into a standoff of conflicting intuitions. In contrast, replies such as the 'phenomenal concept strategy' aim to do justice to the conceivability of zombies and try to explain the conceivability (without being committed to an ontological conclusion). Since Horgan's argument is abductive, its focus is on the conceivability of the CP-zombie (and not on its metaphysical possibility). The conceivability of the scenario calls out for an explanation. Simply denying that the CP-zombie is conceivable by holding that his sensory phenomenology is such that it entails having specific thought contents leaves the conceivability intuition of Horgan's scenario unexplained.

114
Martina Fürst is true that on this view the phenomenology of thought is reduced to sensory phenomenology. However, the sensory phenomenology is highly enriched with cognitive elements (in particular, it is internally structured by cognitive elements). For those who think that there is also non-conceptual sensory phenomenology, holding that the phenomenology of thought is constituted by such conceptually structured sensory phenomenology, might already qualify as a version of liberalism. (I return to this controversy in Section 4.1). Accordingly, most conservatives do not choose interpretation (A). For example, Prinz (2011), Tye andWright (2011), andCarruthers (2014) hold that sensory phenomenology is non-conceptual. To account for the constitution of thoughts, they incorporate a conceptual factor in a different way, for example by focusing on the causal impact of concepts on the non-conceptual sensory phenomenology. 17 (I discuss this view, on which Horgan's scenario is conceivable, in detail in Section 4.2.)

Interpretation (B): Non-conceptual sensory phenomenology (and a missing phenomenal factor)
Another view has it that sensory phenomenology is non-conceptual. That means it consists, for instance, of visual imageries of patterns of lines and color shades, auditory imageries of unparsed sounds and noises and of raw feels. Does this stand in tension to Horgan's similarity claimnamely that the CP-zombie is supposed to have the same sensory phenomenology as his Chinese-understanding twin? No. According to this view, the native Chinese-speaker has non-conceptual sensory phenomenology as well. Thus, conceiving of the CP-zombie as having non-conceptual sensory phenomenology meets the requirement of sensory phenomenal similarity with the Chinese-understanding twin. On the assumption that sensory phenomenology is non-conceptual, Horgan's CP-zombie seems conceivable, at first glance. Suppose that the CP-zombie utters the Chinese words with the content "There is a bottle of wine in the fridge." Behavioral isomorphism (such as the relevant utterance and the behavior of walking to the fridge) and non-conceptual sensory phenomenology (inner visual imagery of patterns of lines and shades, auditory imagery of unparsed sounds) do not entail a content-specific experience of having the thought that there is a bottle of wine in the fridge. However, on a closer look, it turns out that if this were all there is to sensory phenomenology, the Chinese speaker would lack understanding 17 On another view, non-conceptual sensory phenomenology is merely accompanying conscious thoughts. If so, this view does not account for the key issue in Horgan's argumenthow phenomenology constitutes contentand, hence, I set this view aside. Since I am interested in alternative explanations of Horgan's argument, I focus on the strong CP-thesis that aims to account for the constitution of conscious thoughts via phenomenology.
On the Conceivability of a Cognitive Phenomenology Zombie 115 experiences too and, hence, the overall scenario turns out to be inconceivable. Therefore, to make the scenario conceivable, further factors have to be incorporated into the picture that differentiate the phenomenology of the twins. To arrive at a strong liberalism, merely stipulating sensory phenomenology as non-conceptual does not suffice. A further, implicit claim does the work, namely that this is all there is to sensory phenomenology. In particular, the possibility of conceptualization via a cognitive factorfor instance, a causal impact of concepts on the sensory phenomenologyis excluded from the picture. 18 Only the conjunction of these two claims provides us with a way to conceive coherently of the scenario that points toward the existence of a proprietary cognitive phenomenology as the decisive factor which the CP-zombie lacks.
The worry is that this interpretation of the scenario assumes away the problem as well. Construing sensory phenomenology as non-conceptual and denying any conceptual impact on sensory phenomenology makes sensory phenomenology an obvious non-starter for constituting conscious thoughts. This particular reading of nonconceptual sensory phenomenology leads to the intuition that the zombie lacks understanding experiences. However, to make the overall scenario described by Horgan conceivable in this way, it is assumed that (a) sensory phenomenology is non-conceptual and (b) that only an additional phenomenal factor can account for the difference between the Chinese speaker and the zombie. The non-conceptuality of sensory phenomenology alone does not point toward a proprietary cognitive phenomenology, only both assumptions together do. Therefore, on a closer look, this interpretation suffers from the same flaw as the first one. It is construed in such a way that it already adjudicates the issue. Since the aim of the CP-zombie scenario is to argue for the very existence of a proprietary cognitive phenomenology that constitutes conscious thought, one should not (a) present a version of sensory phenomenology which obviously falls short of constituting conscious thoughts and (b) stipulate that the difference in the scenario is best explained by an additional, phenomenal, factor. 19 18 I am grateful to an anonymous referee for drawing my attention to this further claim that is needed to arrive at strong liberalism. 19 Notably, Horgan thinks that he need not rely on non-conceptual sensory phenomenology to make his argument work. Rather, his strategy is the following: (a) Give the guy whatever sensory phenomenology is recognized by the given opponent of cognitive phenomenology, however rich the opponent thinks that sensory phenomenology is. (b) Describe the guy as having all that but not yet understanding Chinese. (c) So it could include, e.g., a sensory state as-of red wine in the fridge, etc., but no Chinese-understanding phenomenology, and lots of spontaneous-seeming desires to move his body in various ways. So far, the situation is the following: for the morph-sequence argument to get off the ground, we need to agree on the target intuition that the CP-zombie is conceivable. Therefore, we have to find a reading of sensory phenomenology that makes the scenario described by Horgan conceivable, but without adjudicating the issue already at the outset. Accordingly, we have to avoid the following readings of sensory phenomenology: (1) A scenario in which the CP-zombie has a sensory phenomenology conceptually structured in such a specific way to be content-individuative and thus to provide him with a specific thought. (Interpretation A) Stipulating such specific conceptual sensory phenomenology would make the scenario described by Horgan inconceivable. This interpretation assumes away the problem at hand. Moreover, it does not provide us with an explanation for the intuition that we can conceive of the CP-zombie as lacking Chinese understanding experiences.
(2) A scenario in which the CP-zombie has non-conceptual sensory phenomenology and the only difference with his Chinese-understanding twin is that the latter also has additional proprietary cognitive phenomenology which is necessary for consciously thinking the target thought. (Interpretation B) Stipulating non-conceptual sensory phenomenology together with the assumption that the CP-zombie lacks an additional phenomenal factor makes the scenario conceivable. However, on this interpretation, the defender of strong liberalism presupposes a picture of sensory phenomenologyas non-conceptual and as not The idea is that we can conceive of the guy as having very rich, sensory-phenomenal, intentional content, but he is simply not phenomenally 'contemplating' that content. I think this way of putting the argument faces two problems. First, on this interpretation, it is hard to see how Horgan's scenario can establish the content-individuative power of cognitive phenomenology. It seems that sensory phenomenology suffices to constitute thought-content, and cognitive phenomenology just contributes to make the former salient to the subject, by contemplating it. Would this turn cognitive phenomenology into just a general feeling of understanding? It is not clear how to exactly understand the hiatus between a phenomenally given thought content that the subject already possesses, and the contemplation of this content. The gist of the issue is that any interpretation that grants that sensory phenomenology accounts for the constitution of thought content will no longer amount to a strong liberalism. Second, and dialectically important, most conservatives would reply that they cannot conceive of what Horgan asks them to. If the guy has the appropriate rich sensory phenomenology, they will think he has the relevant conscious thoughts. Maybe 'contemplating' that content is a further element to having the content itself, but it is still a sensory-phenomenal elementfor example a general feeling of understanding. On their view, sensory phenomenology is all there isthere is no additional, non-sensory, phenomenal element that can be subtracted from the scenario. For this reason, my strategy here is to rather find a description of the scenario that is conceivable for both defenders and opponents of proprietary cognitive phenomenology.
On the Conceivability of a Cognitive Phenomenology Zombie 117 conceptualized by non-phenomenal cognitive factorsthat simply cannot constitute specific thought contents. I analyzed two interpretations of sensory phenomenology that entail the failure/ success of the scenario outlined by Horgan. 20 The lesson to draw is that to evaluate the argument, two readings of sensory phenomenology should be avoidedthe much conceptually enriched reading and the very weak reading of sensory phenomenology. A neutral way of conceiving of Horgan's scenario, one that does not adjudicate the issue at the outset, is desirable. In what follows, I propose a reading that aims at finding a middle ground between the non-conceptual and the conceptually structured sensory phenomenology. To meet this desideratum, I suggest conceiving of the CP-zombie scenario as having sensory phenomenology that is conceptualized, but to the extent that it could be utilized to constitute a range of different contents. Accordingly, such coarse-grained sensory phenomenology would not entail only one specific thought content. What could such conceptualized sensory phenomenology be like? It is a difficult and interesting task to flesh out conceptualized sensory phenomenology in detail. However, my strategy here is to bypass these issues, and to stick with the bare minimum that is needed to develop a neutral way of conceiving of Horgan's scenario.

Interpretation (C): Coarse-grained conceptualized sensory phenomenology
I propose to think of the CP-zombie's sensory phenomenology as involving inner imageries, emotions and bodily movements that exhibit 'openness' to a range of interpretations. 21 Let me clarify what I mean with this qualification of 'openness'. 20 As noted earlier, not all philosophers on the opposite sides of the debate disagree about the characterization of sensory phenomenology. Some conservatives (e.g. Tye and Wright 2011;Carruthers 2014) take sensory phenomenology to be non-conceptual. To account for a particular conscious thought, they add a further factor that interpretation (B) excludes. For example, on Carruthers' view, the non-conceptual sensory phenomenology gets interpreted via the comprehension system and thereby makes a particular content available to the consciousness: There is every reason to think that conceptual information that is activated by interactions between mid-level areas and the association areas (…) gets bound into the content of attended perceptual states and is broadcast along with the latter. Hence (…) we don't just hear a sequence of phonemes when someone speaks, but we hear what they are saying. (Carruthers 2014, 148) 21 I am unsure about how we should conceive of the sensory element of inner speech. To flesh out sensory phenomenology as inner speech posits a special challenge within the debate about cognitive phenomenology. On the one hand, to conceive of inner speech as the sensory act of entertaining a specific content begs the question against liberalism. On the other hand, to conceive of inner speech as just unparsed noise begs the question against conservatism. Perhaps parsed sound elements that are conceived as distinctive items, but not tied to an understanding-experience of a specific meaning, could provide a neutral interpretation of the scenario.

118
Martina Fürst Suppose that someone has inner imageries, emotions and bodily movements that perfectly fit with the target thought that There is a bottle of wine in the fridge, but would also fit with a slightly different thought, e.g. the thought that There is a Vermentino 2015 in the fridge. For example, one has a sensory phenomenology that includes an inner image of a dark bottle with a white label in the fridge, the emotion of looking forward to drinking a glass of good white wine, the bodily movement of walking in the kitchen, and so on. Thus, entertaining the relevant thought does not mean to introduce any additional sensory elements; rather, it means that the coarse-grained conceptualized sensory imagery is pinned downed to one single, meaningful, combination. 22 All the sensory-phenomenal elements that constitute the content of the target thought are already there. But since the sensory phenomenology is coarse-grained, the very same sensory imageries could be involved in a range of different thoughts such as that There is a bottle of Vermentino 2015 in the fridge or that There is alcohol in the kitchen. It is hard to see how sensory phenomenology alone would fix one thought content rather than another. That is why we conceive of the CP-zombie as lacking an experience of the particular target thought. At this point, one might object that this reading already points toward the existence of a proprietary cognitive phenomenology, since on this picture it is a nonsensory phenomenal element that pins down the particular thought content. However, as I will argue in Section 4, there are two explanations at hand and only one of them requires the existence of a proprietary cognitive phenomenology.
To conceive of the sensory phenomenology of the CP-zombie as exhibiting this kind of openness achieves three goals: first, this scenario is close enough to satisfy the requirement of behavioral isomorphism and sensory-phenomenal similarity to a Chinese speaker. Second, the scenario accommodates the intuition that we can coherently conceive of the CP-zombie as lacking an experience of the target thought. Third, this scenario does not assume away the problem at the outset. It neither presupposes that sensory phenomenology is conceptual in such (T)he EQ character of her experience includes (…) her taking her experience to be experience of a certain particular thing. It is this taking, which is part of (…) the cognitive-experiential character or content of her experience, that settles the question, given her causal context, of which of her thought's causal antecedents her thought is aboutin a way that nothing else can.
Even if Strawson is here primarily concerned to offer a solution to the stopping problem tied to causal theories, one could adopt the notion of 'taking as' also to explain how a coarse-grained sensory phenomenology, if taken in a particular way, can yield a determinate content.
On the Conceivability of a Cognitive Phenomenology Zombie 119 a specific way to entail the target thought nor that sensory phenomenology is non-conceptual whichabsent any further cognitive factormust fail to constitute specific thoughts. It only has it that sensory phenomenology is conceptualized and leaves the particular way in which it is conceptualized open. As we will see in Section 4, there are at least two ways to specify the relevant conceptualization.
Horgan's CP-zombie scenario is underspecified with respect to the notion of 'sensory phenomenology.' This notion is crucial for the argument which aims to show that even if the CP-zombie shares the sensory phenomenology with his twin, this does not entail that he has the target thoughts. Therefore, we have to be clear about the notion of sensory phenomenology to evaluate the argument. I think there are at least three ways we can interpret the sensory phenomenology in play. However, only one of them, the one proposed under the label 'interpretation (C)', meets the desideratum of neutrality.
The result of the analysis so far is twofold. First, if we aim to do justice to the scenario outlined by Horgan, 23 and if we do not want to assume away the problem at the outset of the argument, we should conceive of the scenario in the following way: the CP-zombie shares with his twin the externally observable behavior and the sensory phenomenology. Importantly, the latter is conceptualized, but not to the extent that it entails specific thoughts.
Second, the analysis offers an insight that generalizes to most phenomenal contrast arguments: phenomenal contrast scenarios that invoke sensory phenomenology cannot be used to argue for liberalism*. Recall that liberalism comes in two strengths. Liberalism* is the view that the proprietary cognitive phenomenology is modally independent from sensory phenomenology. That means, a person who does not have any sensory phenomenology could still enjoy proprietary cognitive phenomenology. Since sensory phenomenology is a crucial aspect of the CP-zombie scenario, this scenario is ill-suited for establishing liberalism*. 24 According to liberalism+, although cognitive phenomenology depends upon sensory phenomenology, it still qualifies as something over and above sensory phenomenology. Hence, the CP-zombie scenario might still be a promising way to establish liberalism+. In the next section, I will investigate whether Horgan's argument reaches this aim. 23 That is, we should try to meet the similarity requirement in the argument and we should also try to accommodate the intuition that we can conceive of the CP-zombie as lacking Chineseunderstanding experiences. 24 A way of arguing for liberalism* is via arguments from phenomenal contrast that do not involve sensory phenomenology, such as Kriegel's (2015) 'Zoe-Argument ' and Chudnoff's (2015) 'glossed argument' from mathematical intuitions. I discuss the challenges these arguments face in Fürst (2017).

Does the conceivability of the CP-zombie establish strong liberalism?
Does the conceivability of the guy who (a) shares the behavior with a Chinese speaker and (b) has conceptualized sensory phenomenology (with some openness) popping up when uttering Chinese sentences indeed "offer a new argument … in favor of the (distinctive, proprietary, individuative) phenomenology of languageunderstanding" (Horgan 2013, 241)?
To conceive of the guy in this way does not entail that he is consciously thinking a particular thought. According to Horgan, the best explanation for this is that he lacks the proprietary cognitive phenomenology. However, there are two options available to explain why the CP-zombie is not consciously entertaining the target thought. These competing explanations share the view that sensory phenomenology is conceptualized, but they offer different explanations of how this conceptualization takes place. In particular, they disagree about whether concepts have a constitutive or a causal impact on sensory phenomenology. Let me analyze these two options that differ in their breadth and strength.

Strong liberalism
Horgan's CP-zombie argument aims to establish strong liberalism. One explanation of the scenario that qualifies as strong liberalism (more precisely, as strong liberalism+) has it that the CP-zombie is not consciously entertaining the target thought because he lacks the proprietary cognitive phenomenology which is needed to pin down particular thought contents.
One way to defend this interpretation of the scenario is to hold that concepts have a constitutive impact on sensory phenomenology to yield specific thought contents. That means, to consciously entertain a specific thought implies that the right conceptual impact on sensory phenomenology has already taken place. The result is a sensory phenomenology that includes the representations of highlevel properties 25 such as being a bottle of wine or being a fridge. The explanation of the CP-zombie scenario then would be the following: the CP-zombie has coarse-grained conceptualized sensory phenomenology, but he lacks the appropriate high-level sensory phenomenology that is generated and essentially structured by the right conceptualization. Here, the impact of concepts is significant since it is directly built into the sensory phenomenology and essentially modifies the phenomenology that constitutes conscious thoughts.
Some philosophers might object that high-level sensory phenomenology does not count as proprietary cognitive phenomenology. After all, high-level sensory phenomenology is still sensory phenomenology, they might say. For example, 25 For an account of high-level sensory phenomenology, see e.g. Siegel (2010).
On the Conceivability of a Cognitive Phenomenology Zombie 121 Bayne and Montague think that "the label 'cognitive phenomenology' is best reserved for a class of phenomenal properties that are non-sensory" (2011,12). Others, however, think that high-level sensory phenomenology clearly counts as cognitive phenomenology. For example, Montague holds: Since the representation of "high-level" properties can only really be understood in terms of some connection to concepts and the possession of concepts, it seems most accurate to count their associated phenomenology as instances of cognitive phenomenology. (2016,179) Since the question whether high-level sensory phenomenology counts as proprietary cognitive phenomenology is controversial, let me outline another view based on the idea that concepts make a constitutive impact on the phenomenology of conscious thought: one might think that the sensory phenomenology is constituted by concepts in such a way that the resulting phenomenology tied to the content is no longer purely sensory, but rather a new kind of phenomenology, a hybrid of cognitive and sensory phenomenal aspects. 26 The explanation of the CPzombie scenario then would be the following: the CP-zombie has coarse-grained conceptualized sensory phenomenology, but he lacks the appropriate hybrid phenomenology that is fine-grained enough to constitute a particular thought. On this view the impact of concepts is significant since it generates a new, overall, phenomenology. The sensory phenomenology is only one (undetachable) part of the overall phenomenal state and as such it is experienced as meaningful. This view qualifies as strong liberalism+, since the content-individuative phenomenology is a new, hybrid, phenomenology (that is brought about by the phenomenal impact of concepts on the sensory manifold).
Obviously, on both these interpretationsthe high-level sensory view and the hybrid phenomenology viewcognitive phenomenology is not independent from sensory phenomenology. However, even if cognitive phenomenology is intimately tied to sensory phenomenology, it is still something over and above sensory phenomenology. Thus, if this is the best way to explain the conceivability of the CP-zombie, strong liberalism+ is established.

Strong conservatism
Next, let me consider an alternative explanation of the scenario that qualifies as a strong conservatism. One might think that the CP-zombie is not consciously entertaining the target thought because he is not in the position to deploy the relevant concepts. Carruthers and Veillet (2011, 51f.) discuss the distinction between a constitutive and a causal impact of concepts on sensory phenomenology. An explanation of Horgan's scenario can be built on this distinction by holding that the CP-zombie scenario just shows that concepts have a causal impact on the sensory phenomenology which constitutes conscious thoughts. This explanation is available to conservatives and can be developed along the following lines.
First, it is very likely that a defender of strong conservatism will hold that concepts have a causal impact on sensory phenomenology. Why? Recall that strong conservatism has it that sensory phenomenology individuates conscious thoughts. Given the well-known indeterminacy problems of sensory phenomenology, 27 she has to offer some explanation of how concepts figure in the individuation of conscious thoughts. Plausibly, she will add that the sensory phenomenology that individuates conscious thoughts has to be caused by the right concept deployment. For example, Carruthers and Veillet (2017) hold that concepts have a causal impact on sensory phenomenology. They discuss an earlier zombie scenario by Horgan (2011) in which the zombie and his twin are identical also in terms of the concept deployment and they find this scenario inconceivable. In particular, they hold that "it seems impossible that Andy1 and Andy3 could truly be causally/functionally identical (in terms of linguistic processing and concept deployment) while Andy1 experiences a spoken English sentence as meaningful and Andy3 experiences it as meaningless" (2017, 82, emphasis added).
Let me emphasize that in the morph-sequence argument the situation is different. The CP-zombie in the argument developed in Horgan (2013) suffers from a memory deficit which explains why he cannot learn Chinese. This aspect of the argument is crucial for the conceivability of the scenario, conservatives might say. Importantly, the memory deficit of the CP-zombie has a further consequence, namely that he cannot relate specific concepts to his sensory phenomenology tied to Chinese. 28 Thus, the conceptual deficit of the CP-zombie might be due to his memory deficit, rather than due to the alleged lack of a proprietary cognitive phenomenology. In short, conservatives can offer the following explanation of the scenario: given that the causal chain is blocked by the memory deficit, the CPzombie is left with sensory phenomenology in the absence of specific concepts and, hence, he lacks the relevant understanding experiences. 29 The view in the background of the explanation is this. Sensory phenomenology can individuate 27 For a discussion of this problem, see Pitt (2004) and Horgan and Graham (2012). 28 One need not claim that the CP-zombie lacks the relevant concepts. It is very likely that the CP-zombie possesses the relevant concepts, given that he has understanding-experiences in English. The problem is rather that no specific concepts are in play when he is enjoying sensory phenomenology tied to Chinese. 29 At this point, liberals might try to run the thought-experiment without the memory deficit, and thus block the second interpretation. However, the memory deficit seems to be a crucial element to make the CP-zombie conceivable for both liberals and conservatives about cognitive phenomenology.
On the Conceivability of a Cognitive Phenomenology Zombie 123 conscious thoughts only if it is caused in the right way, i.e. by the right concept deployment. The phenomenology of conscious thoughts is all sensory; but it is a sensory phenomenology on which concepts have executed causal power. Let me illustrate this view with an example. Consider the ambiguous duckrabbit figure. On one interpretation, seeing this ambiguous figure counts as a Gestalt-switch phenomenon that results in two different phenomenologiesone which exhibits a specific what-it-is-likeness to see a duck and one which exhibits a specific what-it-is-likeness to see a rabbit. On an alternative interpretation, different concepts are related to the very same sensory phenomenology. The latter interpretation can be used to establish conservatism along the following line: if a person is conceptually primed 30 by bird-stimuli, the priming effect will bring about the deployment of the concept "duck" to her sensory phenomenology. Thus, the sensory phenomenology, caused by the deployment of the "duck" concept, will be utilized to constitute duck-involving thoughts rather than rabbit-involving thoughts. The causal influence of the concept deployment is necessary to pin down the ambiguous sensory phenomenology in a way to constitute either duckor rabbit-thoughts. (Let me emphasize that this reading does not involve a highlevel sensory phenomenology that is constituted by concepts. Rather, the concept deployment is an external element to the sensory phenomenology and does not modify the latter.) 31 On this view, the conceivability of Horgan's CP-zombie scenario is explained by the lack of a non-phenomenal elementthe lack of the concept deployment which is needed for the relevant kind of sensory phenomenology to constitute a specific thought. This qualifies as conservatism about the nature of cognitive phenomenology since the lacking conceptual element (a) need not be phenomenal, and (b) is not an essential part of the phenomenology of conscious thoughts. Moreover, the power of the conceptual element here is only contingently content-individuative, given its causal impact on sensory phenomenology. 32 In principle, another factor that has the right causal impact could do the work. 30 'Priming' is an effect in which exposure to certain stimuli influences the response to a target stimulus. 31 Levine (2011, 110) discusses a further interpretation of the duck-rabbit figure. Cognitive factors might influence our focal attention and fixation points which result in different external stimuli. In this case, the sensory phenomenology of seeing the figure as a rabbit and seeing it as a duck differs, but not due to a direct cognitive penetration but rather due to the effect of different external stimuli. 32 Some might extend this to a thesis on which the cognitive phenomenology depends on both such caused sensory phenomenology and functional roles to end up with determinate thought content. Others (Pautz 2013) think that sensory phenomenology joined with functional roles alone suffice to explain determinate conscious thoughts: "the total intentional facts about a population supervene on the total sensory and (wide and narrow) functional facts about that population, without any need for cognitive phenomenology" (Pautz 2013, 220).

124
Martina Fürst On this interpretation, cognitive phenomenology would neither be independent from sensory phenomenology, nor irreducible to sensory phenomenology. Rather, the phenomenology of conscious thought is identical with sensory phenomenology. The only additional element (which the CP-zombie lacks) is that the latter has to be caused by the right concept deployment. Since the concept deployment need not be seen as a phenomenal element, sensory phenomenology is all the phenomenology we are left with. 33 Thus, this view qualifies as conservatism about the phenomenology of thought.
To sum up: conservatives can explain the conceivability of Horgan's CP-zombie scenario. It suffices to grant (the less-controversial) sensory phenomenology and to add that the right causal impact of concepts on the sensory-phenomenal manifold is crucial. Since the CP-zombie cannot relate specific concepts to his sensory phenomenology, he lacks the conscious target thoughts. The causal influence of concepts on sensory phenomenology can do all the explanatory work needed. Hence, Horgan's scenario does not give us a reason to favor liberalism over the less controversial and parsimonious conservatism about the phenomenology of thought.

Conclusion
We found a way to conceive of Horgan's CP-zombie scenario that meets the desideratum of initial neutrality. On this neutral interpretation, the scenario successfully demonstrates that sensory phenomenology alone does not suffice to constitute specific conscious thoughts. Is the lack of proprietary cognitive phenomenology the best explanation for this intuition and, hence, does the scenario establish a strong liberalism? It does not.
According to strong liberalism, conscious thoughts are constituted by proprietary cognitive phenomenology. First, the scenario does not establish liberalism*, i.e. the view that cognitive phenomenology is modally independent from sensory phenomenology. (Notably, all scenarios that involve sensory phenomenology turn out to be inconclusive in this respect. Thus, liberals* have to search for other arguments in support of their view.) Second, the CP-zombie scenario also does not establish liberalism+, i.e. the view that cognitive phenomenology is irreducible to sensory phenomenology. The reason is that the conceivability of the 33 Conservatives who might adopt that explanation are, for example, Carruthers and Veillet. They hold that "when one understands a sentence one is aware, both of its sensory properties (pitch, loudness, phonology, and so forth) and its semantic ones. … So, yes, the experience of a sentence as meaning that ducks are cute is a phenomenal state; it doesn't follow that the meaning-component of that experience has phenomenal properties, nor that it makes a constitutive, irreducible, contribution to the phenomenal properties of the entire hearing-event" (2017, 83).
On the Conceivability of a Cognitive Phenomenology Zombie 125 scenario can be explained equally well by a conservative view that only grants sensory phenomenology and adds that sensory phenomenology has to be caused by the right concepts to constitute conscious thoughts. The conceivability of Horgan's scenario might be due to the guy's memory deficit and failure to relate specific concepts to sensory phenomenology, rather than due to the lack of a proprietary cognitive phenomenology. Thus, the conceivability of a CP-zombie fails to settle the debate about whether proprietary cognitive phenomenology or sensory phenomenology is constitutive of conscious thoughts. The roots of the ineffectiveness of the argument lie in the diverse readings of sensory phenomenology that flesh out the relation between sensory phenomenology and concepts differently, but explain the conceivability of the CP-zombie equally well. The lesson to be learned is that we cannot adjudicate the debate about cognitive phenomenology, unless we have clarified the notion of sensory phenomenology first.*