Consciousness and Cognition 56 (2017) 1–12

Contents lists available at ScienceDirect

Consciousness and Cognition

journal homepage: www.elsevier.com/locate/concog

Unitary and dual models of phenomenal consciousness MARK

a,⁎ b
Tomáš Marvan , Michal Polák
a
Institute of Philosophy of the Czech Academy of Sciences, Jilská 1, Prague 110 00, Czech Republic
b
Department of Philosophy, University of West Bohemia, Pilsen, Czech Republic

AB S T R A CT

There is almost unanimous consensus among the theorists of consciousness that the phenomenal
character of a mental state cannot exist without consciousness. We argue for a reappraisal of this
consensus. We distinguish two models of phenomenal consciousness: unitary and dual. Unitary
model takes the production of a phenomenal quality and it’s becoming conscious to be one and
the same thing. The dual model, which we advocate in this paper, distinguishes the process in
which the phenomenal quality is formed from the process that makes this quality conscious. We
put forward a conceptual, methodological, neuropsychological and neural argument for the dual
model. These arguments are independent but provide mutual support to each other. Together,
they strongly support the dual model of phenomenal consciousness and the concomitant idea of
unconscious mental qualities. The dual view is thus, we submit, a hypothesis worthy of further
probing and development.

1. Introduction

There is almost unanimous consensus in the literature regarding consciousness that the phenomenal character of a mental state
exists exclusively in its conscious form (see, e.g., Revonsuo, 2010, 39 and 72, Crick & Koch, 2003; Koch, 2004, 242 and 300, ffytche,
2000; Prinz, 2012). Pains, sounds and colours – phenomenal qualities constituting the phenomenal character of a mental state – can
only occur at the level of consciousness. Sophisticated neural processing accompanies these mental states but this processing entirely
lacks phenomenal character. We argue for a reappraisal of this consensus. We distinguish two models of phenomenal consciousness:
unitary and dual. The unitary model takes the production of a phenomenal character, and its becoming conscious, to be one and the
same thing. Most theories of consciousness advocated in the philosophy of mind and cognitive neuroscience of consciousness are
unitary. The dual model, which we describe and defend in this paper, distinguishes the process in which the phenomenal quality is
formed from the process that makes this quality become conscious. The dual model agrees with the unitary model in that conscious
phenomenal states possess the experiential what-it’s-like-ness and that unconscious mental states lack it. However, it allows that
unconscious mental states possess phenomenal character (or phenomenality, as Ned Block calls it; see Block, 2001). On this view, the
experiential what-it’s-like-ness is simply a result of the combination of phenomenal character and consciousness.
It is important to realize that the widely held identification of the experiential what-it’s-like-ness with phenomenal character, and
the ensuing denial of unconscious phenomenality, is not a result of evidence-based argument. Rather, the notion of unconscious
phenomenal quality seems to be a contradictio in adjecto because phenomenality seems to be conceptually linked with consciousness.
We propose to sever this alleged conceptual link. The stubborn intuition that only conscious mental states can harbor phenomenal
character is probably based on the fact that we come to know the phenomenal character of mental states from the instances of
conscious perception and other kinds of conscious mental processes. But that does not necessarily imply that phenomenal character

⁎
Corresponding author.
E-mail address: marvan@flu.cas.cz (T. Marvan).

http://dx.doi.org/10.1016/j.concog.2017.09.006
Received 15 March 2017; Received in revised form 13 September 2017; Accepted 13 September 2017
1053-8100/ © 2017 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

Fig. 1. The dual model of phenomenal consciousness distinguishes the process in which the phenomenal character of a mental state is formed from the process that
makes this character conscious. Thus the phenomenal character can occur unconsciously and the what-it’s-like-ness of a mental state is the phenomenal character as
experienced consciously by a subject.

could not exist without consciousness. And neither does it imply that we can determine the presence of the phenomenal character of a
mental state only by becoming conscious of that state. Our proposal, schematically depicted in Fig. 1, is revisionary but, as will
become clear, supported by a number of compelling arguments.
Many authors claim that episodes of conscious and unconscious perception are of a very similar kind. According to Jesse Prinz, for
example, unconscious vision is “very much like” conscious vision (Prinz, 2015, 381). Yet he rejects the notion of unconscious
phenomenal qualities: the difference between conscious and unconscious perceptual states is, on his account, precisely that the first
have phenomenal character whereas the latter don’t (see also Prinz, 2012, esp. pp. 143–145). Phenomenal character, for him, is what
distinguishes conscious and unconscious perception. We disagree: phenomenality can be unconscious and the episodes of conscious
and unconscious perception are even more similar then Prinz is ready to admit. In the following sections, we state the case for
unconscious phenomenality, adduce empirical evidence in its support and draw the consequences for the construction of empirically
sound theories of consciousness. We argue for the possibility of unconscious phenomenal character – and for the associated dual
model of phenomenal consciousness – by means of four distinct yet mutually supportive arguments: conceptual, methodological,
neuropsychological and neural arguments.

2. Conscious and unconscious phenomenal qualities (I): Conceptual and methodological arguments

2.1. The conceptual argument

What we do in this paper can be seen as a reinforcement and expansion of the argument for unconscious phenomenality put
forward by David Rosenthal (see esp. Rosenthal, 2005a, chs. 5–7, and Rosenthal, 2010). We should like to stress that we do not
subscribe to his higher-order thought (HOT) theory of consciousness (for which see Rosenthal, 1997). According to the HOT theory of
consciousness, a mental state is made conscious by a distinct thought with a suitable content directed towards it. The HOT theory has
been subjected to what we tend to believe are fatal criticisms, both of a conceptual (see, e.g., Block, 2011; Byrne, 1997) and empirical
nature (Kozuch, 2014). But Rosenthal’s central insight that a mental state can retain its phenomenal character even when un-
conscious is independent of the HOT theoretical framework, as Rosenthal himself acknowledges (Rosenthal, 2011, 435n.7). When
taken as separate from the HOT framework it is, we submit, a vital standalone hypothesis deserving of further probing and devel-
opment.
Although we reject the HOT account of how mental states become conscious, we share Rosenthal’s view that a mental state’s being
conscious is an extrinsic property of that state and that unconscious mental states can be intrinsically the same as their conscious
counterparts. However, in contrast to Rosenthal, we propose a neural explanation of how a state becomes conscious (see Section 3.2).
We are not persuaded that a psychological style of explanation is required, as Rosenthal (2012) insists. Instead, we propose a bottom-
up account of phenomenal consciousness in which the relevant neural mechanisms do all the explanatory work; that is, they enable to
explain the difference between conscious and unconscious mental states. According to the dual model, consciousness does not really
modify the intrinsic phenomenal properties of a state. We can see, hear, smell, etc., without awareness. As Prinz (2017) remarks, we
can do so many things without consciousness that the question “What is consciousness actually good for?” becomes difficult to
answer. It seems that its proper function is to make perceptual contents available for flexible control of behaviour and for verbal
reports. This fits well with the dual model of consciousness.
Now, many philosophers believe that it makes no sense to say that, for instance, a pain could occur unconsciously. The only way
for a pain to exist is for it to be consciously felt. The distinction between appearance and reality seems to collapse in the case of pains.
This prevailing view, endorsed, e.g., by Saul Kripke (1980) and Fred Dretske (2006), was expressed already at the beginning of the
20th century by the then significant philosopher Alexander Bain; he wrote of pleasure that it is no pleasure, if not conscious (Bain,
1894, 352). Rosenthal breaks free of this tradition. He suggests that consciousness is something that phenomenal states can but need
not have.
Rosenthal (1986) illustrates the distinction between conscious and unconscious pain in the following way. Suppose we are in pain
for a longer period of time, say the whole day. Although we are not conscious of the pain for all this time, being often distracted by
other things, it is perfectly natural to say that we have been in pain the whole day. By the same token, someone may limp while

2
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

walking without being conscious of the mild pain in the leg (Rosenthal, 2011, 435n8).1 It seems odd to say that the pain was there
only in those intervals when we registered it consciously. Yet this is precisely what the aforementioned philosophical tradition
demands. The tradition dictates that the pain ceases to exist the moment we stop being aware of it, although the same bodily
processes presumably continue to operate all along. Rosenthal prefers to say that the pain, including its phenomenal quality, can
become temporarily unconscious. At the very least, he points out, there are no independent reasons for preferring the traditional story
to the one that allows for unconscious phenomenality. Rosenthal also submits that instances of subliminal perception (Kihlstrom,
1987; Merikle, Smilek, & Eastwood, 2000) prove his point regarding unconscious sensory qualities (Rosenthal, 2011, 433). When a
perception is subliminal, the subject is phenomenally unaware of the percept, but the phenomenal quality persists and is registered
unconsciously. At least this is one way to take it, and it is not obviously less plausible when compared to the standard story, which
denies all unconscious phenomenality.
The Rosenthalian conceptual argument thus states that it is no more unnatural or unfounded to speak about unconscious pains
and other phenomenal states, than it is to hold that phenomenal qualities are necessarily tied with consciousness. It is question-
begging to hold that phenomenal qualities must be conscious – an artifact of the deeply entrenched view that it makes no sense to
speak about phenomenal states apart from their conscious presentations. Although the qualitative nature of a mental state can be
directly ascertained only from its conscious appearance to a subject, consciousness is not constitutive of it. However, it is natural to
worry that because consciousness provides access to the quality’s intrinsic nature, phenomenal qualities are simply consciousness-
making qualities; the idea of unconscious phenomenal states might then strike many theorists as incoherent. There is thus a pressing
need to provide consciousness-independent criteria of identity for perceptual phenomenal qualities.
Quality-Space Theory (QST; Clark, 2000; Rosenthal, 2010, ch. 1) is meant to do precisely this. It fixes perceptual qualities by their
perceptual roles. Perceptual states enable us to distinguish between various properties of perceptible objects because they track these
properties by means of the qualitative features they have. There is thus a matching between the quality space of mental qualities (in a
specified sensory modality) and the space of perceptible properties of objects. Perceptual properties of objects are commonsense
properties such as shape, colour, position or movement of physical things in our surroundings. If we are able to make a perceptual
discrimination between two such perceptible properties, say between two shades of the same colour, that means there is a difference
between mental qualities corresponding to these properties. However, the matching of perceptual properties and mental qualities is
not a matching of any qualitative likeness between the two. Rather, what is analogous are the relations of similarity and difference
between the members of both these groups. Perceptible properties of objects are similar and different in certain respects, and mental
qualities in a given sensory modality display the same relational ordering. If this were not the case, we couldn’t use the mental
qualities to successfully track the perceptible features of external objects and accurately differentiate between them. To give an
example, perceptible red colour of an external physical object is more similar to perceptible orange than to perceptible violet. The
same set of relations of similarity and difference then holds between the corresponding mental qualities as well: the mental quality of
red is more similar to the mental quality of orange than to the mental quality of violet. Both quality spaces are in this sense
homomorphic (see Rosenthal 2005b, Section 4, for more details).
Perceptual states enable perceptual discriminations independently of being conscious: there are various methods of constructing a
quality space of, say, colours, which do not depend on conscious awareness of perceptual qualities. Unconscious priming and forced-
choice discrimination in absence of conscious perception are prominent among these. These empirical methods enable us to de-
termine the relations of relative discriminability and similarity amongst the qualities, and to construct a systematic quality space
reflecting these relations. QST complements the first person access to perceptual qualities by empirically fixing their relative positions
in a quality space without relying on their conscious appearance. The first person conscious access is thus not the only mode of access
to perceptual qualities, although other modes of access are only indirect, via observable behaviour of the tested subject, neuro-
physiological activity measured by imaging methods and other physiological measures such as changes in skin conductance. Again,
even though QST is compatible with the HOT theory of consciousness, it is dissociable from it. There are other ways to account for
how mental states become conscious (we offer a distinctly neural explanation in Section 3.2 below).
Why is QST supporting the idea of unconscious phenomenal qualities? Because perceptual discrimination of stimuli is performed
by means of discriminating the phenomenal features, both at conscious and unconscious level.2 When a subject indicates she can dis-
tinguish two perceptual qualities, she distinguishes them by their phenomenal appearance. And because the same perceptual dis-
criminations within a given quality space can be performed both consciously and unconsciously, we have a reason to believe that
both types of discriminations are performed with the help of the same phenomenal qualities. If, on the contrary, conscious and
unconscious perceptual discriminations (and thus quality spaces) would differ, one could assume that either there are no unconscious
qualities or perhaps that unconscious qualities are in their nature different from the conscious ones. We know of no empirical work in
psychophysics proving that conscious and unconscious quality spaces differ. By contrast with this parsimonious explanation of how
perceptual discrimination is performed, unitary theories, committed to denying unconscious phenomenal states altogether, must hold
that albeit the resulting perceptual discriminations are identical, unconscious discriminations are performed in a completely different
way, with the help of a completely different set of mechanisms than the conscious ones. This is a logical consequence of the fact that
according to the unitary model, no phenomenality is involved in unconscious perceptual discrimination. Unfortunately, unitary
theories leave those alternative mechanisms unspecified.

1
Neuman (2004) reports a similar case in which the patient’s face during surgery had an expression of pain though the patient was under local anesthesia, claimed
to have no subjective awareness of pain and was genuinely surprised by the surgeon’s question “Do you feel any pain?”
2
A parallel argument can be run for perceptual recognition of object properties via phenomenal features. We show this in Section 3.1.

3
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

Rosenthal’s example of unconscious pain opens up the possibility of conceptually separating mental state itself, including its
phenomenal character, from its conscious appearance to the subject. But the debate about unconscious phenomenality need not be
resolved by conceptual intuitions alone. If that were the case, we could probably never reach a resolution of the issue. Fortunately,
there are other arguments in favour of separating the phenomenal character of a mental state from its becoming conscious. We will
develop them in the following sections of this paper.

2.2. Methodological argument

We are well aware that the idea of unconscious phenomenality is counterintuitive. The popularity of the idea of unconscious
phenomenal states among theorists of consciousness is thus negligible. Apart from Rosenthal and some of his collaborators (see
Young, Keller, & Rosenthal, 2014), to our best knowledge the only philosopher advocating unconscious qualitative states is Coleman
(unpublished results).3 He argues that whereas phenomenal character of mental states differs from case to case, the way in which
these states are conscious is always the same. Therefore, it seems plausible to treat the phenomenality and the consciousness of a
mental state as two separate things. Coleman also claims, in a similar vein to Rosenthal, that there are no arguments for the necessity
of conscious nature of phenomenal qualities; there is just the stubborn intuition that they must be conscious.4
To further enhance the attractiveness of the idea of unconscious phenomenality, we propose the following methodological ar-
gument in its favour. In order for consciousness to be amenable to scientific exploration, it needs to be treated as an experimental
variable (Baars, 1997, 2003). This is a standard methodological desideratum in all areas of scientific research. Applied to the case of
phenomenal consciousness: we need to empirically study pairs of mental states that are alike in all respects except that one is
conscious and the other not. Only then will consciousness be treated purely as a variable. This, though, is not how the aforementioned
philosophical tradition conceives of the matter. According to the tradition, conscious and unconscious mental states are quite dif-
ferent. The conscious states have phenomenal character while the unconscious don’t. Yet if unconscious states lack phenomenal
character, their consciousness is not an experimental variable. Rather, we are dealing with mental states being conscious and having
phenomenal character on the one hand, and states lacking both consciousness and phenomenal character on the other. If this were the
case, the possibility of a truly scientific study of consciousness would be endangered.
Embracing the notion of unconscious phenomenality removes this obstacle to the scientific study of consciousness. The dual
model of phenomenal consciousness allows one to say that conscious and unconscious mental states share their phenomenal char-
acter. Supposing this phenomenal character to be fixed independently of consciousness, we can then launch an empirical research
into how a mental state becomes conscious during perceptual and other processes. The study of how a state acquires its phenomenal
character should target other, though probably closely related, brain mechanisms.5
We should like to stress again that according to the dual model, conscious and unconscious states may share the same phenomenal
character. We do not wish to claim that unconscious phenomenal states possess some mysterious “phenomenality” of a hitherto
unknown kind (such as, for instance, the phenomenality of electrons posited by some panpsychists – see Goff, 2016). The consciously
perceived shape, colour and other properties are phenomenal qualities and the same qualities can be realized in the brain even
without consciousness. Conceived this way, the methodological desideratum of empirical research into consciousness is met. We can
treat consciousness of a mental state as a variable that some states exhibit and some other states, possessing the same phenomenality,
lack.6

3. Conscious and unconscious phenomenal qualities (II): Arguments from neuropsychology and cognitive neuroscience

3.1. Neuropsychology

Important empirical support for a dissociation of consciousness and phenomenal character is furnished by a visual dysfunction
called hemispatial neglect (henceforward “neglect”; see Vallar & Perani, 1986). Neglect results most frequently from stroke in the
inferior parietal lobe (particularly in the region of the temporo-parietal junction), but it can also follow damage to other parts of brain
both in the cortex (the inferior frontal cortex and the superior temporal gyrus) and in subcortical areas (the basal ganglia and the
thalamus) (Arend, Rafal, & Ward, 2008; Husain & Rorden, 2003). A person suffering from this disorder cannot see objects in the left
half of her visual field. For example, if a patient with neglect draws a clock hanging on the wall, a house or a cat, the result is that
almost the entire left half of the image is absent. The spectrum of behavioural disorders stemming from this type of neglect is broad:
patients bump into the left side of the door, do not eat food on the left side of the plate, do not register when someone is approaching

3
The neuroscientist Damasio (1999) also hints at unconscious phenomenal states in his talk of “unconscious images”.
4
Referee suggests it is possible to hold that phenomenal consciousness is a determinable of which particular conscious qualitative states are determinates. On this
picture, mental qualities cannot become unconscious. This way of looking at things presupposes the unitary view of consciousness, and in this respect begs the question
against the dual model. Still, it has the merit of showing that there might be unitary conceptions of consciousness that do not merely rely on an unsupported intuition
that phenomenal qualities must necessarily be conscious.
5
We will say more about the possible neural underpinning of the dual model of phenomenal consciousness in Section 3.2 below.
6
According to some contemporary theorists, cognitive mental states such as thoughts or judgements lack all distinctive phenomenality (see, for instance, Carruthers,
2010). So one might object that the methodological argument of this section does not target such conceived states, because they are intrinsically the same in both
conscious and unconscious form. We grant that if thoughts lack phenomenality, their consciousness can be treated as a variable and the methodological desideratum is
satisfied. However, here we focus on perceptual states and these have phenomenality on virtually every account. Moreover, we favour the conception of Prinz (2012)
according to which even cognitive mental states possess phenomenal character; in the case of verbalized thoughts, the phenomenology is that of the subvocal speech.

4
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

from the left, and the like.

For our purposes, the most intriguing thing about neglect is the fact that the absence of conscious phenomenal information in the
left part of the visual field does not preclude the brain unconsciously processing the stimuli from this area. Take perhaps the most
famous example of neglect, an experiment involving pictures of two houses that were shown to a person affected by this visual
disorder (Marshall & Halligan, 1988). The first picture displayed a house with two windows and a chimney, all in green contours. In
the second picture the same house was drawn, but the left half of it, the one that corresponded to the area of the visual field in which
the person did not consciously see anything, was in red contoured flames. The subject was asked two questions: what she saw in the
pictures and whether she could see any difference between them. After answering that she saw houses and that the two pictures were
the same, the task was to choose from the two houses the one in which she would prefer to live. After initial embarrassment the
subject chose one or the other. Although she did not see a burning house, as is evident from the answer to the previous questions, in
repeated trials she selected in about 80% of cases the house without flames. Thus, the flames were somehow registered by the subject,
albeit not consciously.
In our view, the burning house was perceived unconsciously in the following way: the phenomenal image of the house was
formed, but it was not made conscious. This unconscious phenomenal image of the flames engulfing the second house then guided the
subject’s choice of the house she would prefer to live in. We thus believe that in order to account for such a significant behavioural
effect, it is plausible to assume that the brain works with unconscious phenomenal information that tips the subject off to the obvious
choice of the house. Therefore, we take the results of the burning house experiment as providing further empirical support for the
distinction between phenomenal qualities and their conscious awareness. The Quality-Space Theory mentioned in Section 2.1.
stresses that unconscious discrimination of stimuli relies on the same qualitative features as are perceived consciously. Here, the same
idea is in play, only the subject is engaged in unconscious recognition of a stimulus (i.e., in unconscious recognition of the shape of the
flames engulfing the second house).
Some authors debate whether the case of the burning house experiment confirms that patients identified unconsciously the
relevant meaning of neglected stimuli (“It’s a burning house!”), or were only able to process the purely perceptual information such as
the shape of the depicted flames (see Bisiach & Rusconi, 1990; Vallar, Rusconi, & Bisiach, 1994). The level of processing of the
consciously unperceived visual stimuli has been studied for instance by Berti and Rizzolatti (1992) in a study on a single patient. The
patient, affected by left visual field extinction (a type of neglect in which the presence of the stimulus in the ipsilateral hemifield
prevents the patient from seeing the stimulus in the contralateral hemifield), was able to identify and name consciously unperceived
objects depicted in the photograph. As further shown by Berti (2002), unconscious processing of perceptual stimuli can be very
comprehensive. But even if we assume that in the Marshall and Halligan experiment, only the purely perceptual information about
the depicted house was processed, it is still completely consistent with the idea that phenomenal character was perceived un-
consciously and guided the subsequent behaviour of the patient.
According to one influential view, neglect is strictly speaking not a visual dysfunction but rather a kind of attentional dysfunction
(see, e.g., Gazzaniga, Ivry, & Mangun, 2014, 278). Clearly, this view in no way clashes with the dual model of phenomenal con-
sciousness. The so called ventral stream (Goodale & Milner, 1992) where the unconscious phenomenal qualities are presumably
formed remains intact in neglect patients. Attentional dysfunction in the parietal areas then blocks these qualities from becoming
conscious. The case of blindsight, another neuropsychological disorder frequently used as an example of unconscious perception, is
not that straightforward, though. Blindsight is typically caused by lesion to the striate cortex (V1). Due to the lesion, patients cannot
consciously see objects in part of their visual field (scotoma). Despite the impairment of the visual cortex they often perform re-
markably well within a forced-choice paradigm, discriminating simple stimuli such as the location of the object, its space orientation,
direction of movement, or wavelength (Cowey, 2010; Sahraie, Hibbard, Trevethan, Ritchie, & Wieskrantz, 2010).
Suppose that V1 is necessary (though not sufficient) for phenomenal character to be formed. Then, due to lesioned V1, blindsight
patients will not be able to register visual phenomenal qualities, not even unconsciously. Neurophysiological research, however,
indicates that there are alternative neural routes from the retina to various regions of the visual brain; e.g., some routes lead through
the LGN and superior colliculus, bypass V1 and reach extrastriate regions such as V2, V4, V5 (Cowey, 2010; Pöppel, Held, & Frost,
1973; Weiskrantz, 1996). This finding convinced some authors that in blindsight, some residual visual information enters the ex-
trastriate visual regions, and is processed there even without the contribution of V1 (Overgaard, 2012, 609; Weiskrantz, 1986, 1996).
Is it possible that these extrastriate regions are able to process unconscious phenomenal character of visual percepts, despite the
impaired connection between the retina, V1 and extrastriate visual regions? We submit that this is an intriguing hypothesis, which
sits well with the dual construal of phenomenal consciousness. Similarly as in the case of neglect, the above chance correct visual
discriminations of a blindsight subject entitles one to infer that unconscious phenomenal character has been constituted and tacitly
used as a guidance in the discriminatory task at hand.
Still, some researchers believe that blindsight is a highly controversial example of unconscious perception. A hotly debated issue,
for example, is whether the patients’ visual perception during blindsight is totally unconscious (Brogaard, 2011; Weiskrantz, 2009),
or whether it is conscious, though qualitatively degraded and transformed in its own way (Overgaard, Fehl,
Mouridsen, & Cleeremans, 2008; Overgaard & Grünbaum, 2011; Phillips, 2016). Particularly extended and deep criticism can be
found in Campion, Latto, and Smith (1983), where the view of blindsight as phenomenally unconscious seeing was challenged. They
claim that all blindsight subjects are aware of stimuli at least to some extent. Conscious vision in blindsight could depend on islands of
spared V1 or on how the light from the “unseen” stimulus might be detected in the portion of the visual field unaffected by the
scotoma. Its presence may be obscured by the fact that blindsight patients use different response criteria during signal detection in the
blind and in the seeing field. Particularly in the yes/no task, it’s possible that the response depends not only on sensitivity to the target,
but also on a tendency to select the conservative response “No” when subject’s reasons for saying “Yes” are not absolutely persuasive

5
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

(this is the so called ʻconservative response bias’). If this criticism is correct, all blindsight subjects see the stimuli consciously, and this
is so however limited, degraded, or distorted their conscious seeing is (see also Phillips, 2016).
However, without a clearly defined criterion for distinguishing conscious from unconscious visual perception, the debate cannot
be resolved. The often used dichotomic “Seen-Not seen” scale may be too crude to eliminate the conservative response bias. But when
more nuanced subjective measures of consciousness are applied, this danger can be averted. Perceptual awareness scale (PAS;
Sandberg & Overgaard, 2015; Sandberg, Timmermans, Overgaard, & Cleeremans, 2010) is such a measure. Its four-point scale – (1)
‘No experience’, (2) ‘Brief glimpse’, (3) ‘Almost clear image’, and (4) ‘Absolutely clear image’ is quite sensitive to all sorts of ex-
periences the subject might undergo and it was built with the input from the experimental subjects. When PAS is used, some stimuli
classified as unseen on a dichotomous scale are classified as somewhat consciously apprehended (point 2 of the perceptual awareness
scale). The effect of having more categories of experience on a scale instead of just two is that the responses are more precise and the
risk of conservative response bias is minimized. PAS is very sensitive particularly for low stimulus intensities and it has stronger
performance-awareness correlation than objective measures such as confidence ratings or post-decision wagering (Sandberg et al.,
2010). The introduction of the PAS, though, does not eliminate unconscious perception. As Mazzi, Bagattini, and Savazzi (2016)
reports, when PAS was applied on a patient with completely destroyed and dysfunctional left part of V1, only the values of 1 (No
experience) and 2 (Brief glimpse) on a perceptual scale were selected, while performance remained at above chance levels.7 Thus,
blindsight cannot be excluded as a form of unconscious visual perception and may be legitimately used in favour of the dual model of
phenomenal consciousness, until a decisive argument to the contrary is put forward.
Cases of unconscious visual perception in which healthy subjects are tested could, in a similar vein, be used in support of the dual
model of phenomenal consciousness. We have in mind the various experimentally studied forms of subliminal perception (for a
review, see Breitmeyer, 2015). These could be divided into experimental paradigms in which the time to perceive a stimulus is
shorter than the conscious threshold, e.g., in classical visual masking, and into experiments like continuous flash suppression or
binocular rivalry where the time frame is much longer. In both groups, the above chance correct discriminatory reactions of subjects
can be used to infer the presence of unconscious phenomenal qualities. Although unconscious perception in healthy subjects is
considered as a well proven phenomenon (see Dehaene, 2014; Prinz, 2017), some researchers recently raised doubts about it. For
instance, Peters and Lau (2015) deny the reality of forward- and backward-masked unconscious perception. Using their bias-free
experimental design, they conclude that discriminatory success goes hand in hand with conscious awareness, and claim there is no
compelling evidence for unconscious perception. The thresholds for subjective awareness and objective discrimination of stimuli are,
according to them, effectively the same. They claim that unconscious perception doesn’t occur in people without brain damage (see
also Block & Phillips, 2016).
Peters and Lau concentrate on masking experiments involving healthy subjects. By focusing on brain-damaged subjects with
neglect or blindsight, we can largely steer clear of this so far unresolved issue.8 It is possible that the sceptical conclusions of Peters
and Lau are due to limitations of the experimental procedures applied in experiments with healthy subjects. Perhaps the masking
methods they rely on are not sufficient to completely eliminate conscious perception. An advantage of pathological cases is that there
is a precisely localised destruction of brain tissue. Knowing that a certain visual area is dysfunctional allows us to rule out any
processing of the visual information in this part of the brain. The risk of conservative response bias in these cases is correspondingly
reduced: because the subject can form no percepts in the damaged area, conservative response bias can hardly arise in these cases.
And if any suspicion remains that conservative response bias is in place, due perhaps to residual detection of stimuli in undamaged
areas of visual cortex, we can minimize it by applying the PAS method described above.
We would, of course, welcome the extension of our argument for unconscious phenomenality from experiments with neu-
ropsychological patients to experiments with healthy subjects. There is some hope that this could be done. Ro, Singhal, and Garcia
(2009) point out that perceptual experiments on healthy and neurological subjects provide converging evidence that processing of
shapes and other features occurs in the absence of awareness. Still, how smoothly the findings from neuropsychological deficits
generalize to neurologically healthy humans remains an open and somewhat controversial issue. Apart from the scepticism of Peters
and Lau, some experimenters suggest that visual processing in blindsight patients works differently than in healthy subjects, possibly
due to neuroplasticity (Railo, Andersson, Kaasinen, Laine, & Koivisto, 2014; Tamietto et al., 2010). Therefore the safer option (for
now) seems to be to treat neuropsychological perception and ordinary perception separately. Both can support the dual model of
consciousness in their own right, especially if the current controversies are resolved in favour of the unconscious perception in
healthy subjects.9
Now, the thesis that conscious and unconscious states share the same phenomenal qualities might be deemed too strong. It might
be necessary to modify it in light of recent empirical findings. Moreover, some of these findings suggest that some phenomenal
qualities cannot occur without consciousness. Take, for example, the research of Moutoussis and Zeki (2002). These authors used the
dichoptic masking paradigm to study unconscious face perception. First, they exposed subjects to two coloured drawings of faces

7
The Brief glimpse (option 2 on the scale) simply means “I feel that something has been shown, but I have no idea what it was.” The subject is not aware of any
specific details of the stimulus (its colour, shape and so on). It is plausible to assume that successful performance is driven by unconscious processes when option 2 is
selected.
8
Admittedly, some arguments in Phillips (2016) do target perception in neglect and blindsight. We are not persuaded by Phillips’ reasoning, but we do not want to
go into details in this paper.
9
Peters and Lau acknowledge that subliminal priming methods are not affected by their arguments against unconscious perception. Their criticisms target forced-
choice discrimination paradigm (Peters & Lau., 2015, 10). The results of the priming experiments, such as the ones we use to support the dual model in the following
paragraphs, are thus unaffected by current controversies about unconscious perception.

6
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

framed in a square form. One eye was presented with a drawing of a red face with green hair and green face contours, the other with
the colour-opposite drawing, that being a picture of a green face with red hair and red face contours. The resulting conscious visual
percept was that of a yellow square with no face consciously visible. Subsequently, they exposed both eyes with a same-coloured face
(red face, green hair and green contours). This led to conscious visual percept of the face. A comparison between the opposite-colour
stimulation with the same-colour stimulation exposed that the brain regions activated were much the same in both cases. However,
the opposite stimulation led to a weaker activation and a less extensive neural pattern in the fusiform area. Suppose that the
unconscious phenomenal character posited by the dual model of consciousness exists. Could the results of Moutoussis and Zeki be
taken to mean that unconscious phenomenal character of a perceptual state is weaker/less pronounced then the conscious phe-
nomenal character?
If we opt for an affirmative answer, the thesis of unconscious phenomenality will need to be modified accordingly. We will have to
say that the unconscious phenomenality is in fact only a “proto-phenomenality”, a less vivid version of the full-blown conscious
phenomenality. We do not want to reject this proposal outright. In fact, it could even help to explain the uncertainty and degraded
precision typical of recognition tasks during unconscious perception. However, there is an alternative explanation of the etiolated
signal during unconscious face perception: the stronger signal during conscious perception of a face is entirely due to the neural
mechanisms of consciousness, not of phenomenality. This explanation allows one to defend the thesis that unconscious and conscious
phenomenal character are the same.
A more pressing issue for the dual model of phenomenal consciousness is constituted by certain perceptual masking experiments.
Subliminal perceptual priming and forced-choice discrimination work when the shapes of letters of the alphabet, faces, flames
consuming the house, orientation of gratings, etc. are unconsciously perceived (He, Cavanagh, & Intriligator, 1996; Prinz, 2015;
Weiskrantz, 1986). Phenomenal shapes of various kinds thus can be argued to exist unconsciously. However, some experiments
suggest that some phenomenal qualities do not exist during unconscious perception. For example, colours pose a challenge for the
dual model.
The level to which the visual information is processed without consciousness was studied by Breitmeyer, Ro, and Singhal (2004).
They describe a priming experiment involving white and desaturated blue and green stimuli. The subjects were first shown a white
disk for such a short time interval that they could not see the disk consciously (14.3 ms). The white disk served as a prime. Then a
blank screen was presented (28.6 ms), followed by a mask consisting of desaturated green annulus surrounding the area formerly
occupied by the white disk (28.6 ms). The subjects were then asked, in 2 AFC task, whether the annulus was blue or green. The
answers were then compared to the similar situation in which white prime disk was followed by a blank screen which was followed
by a desaturated blue annulus of the same position and size as the green one. It was found that the priming effect of the unseen white
disk depends on the colour of the annulus. When the annulus was green, the white prime raised the choice in favour of the green
annulus. On the contrary, when the annulus was blue, the white disk did not serve as an effective prime for preferring the blue
annulus. Thus the white prime disk somehow facilitates the choice of the green colour; it acts more like a green prime than a blue
prime.
However, when in a variant of the experiment the white prime disks were consciously visible, they acted more like blue primes
than green ones. What accounts for such a striking difference in discriminatory reactions? Breitmeyer and colleagues propose to
distinguish between stimulus-level priming and perceptual-level priming. They speculate that the visual stimulus-level priming acts
solely on the physical information regarding the wavelength properties of the stimulus. Photometric measurements of the display
used in their lab revealed that the wavelength composition of desaturated green phosphors was more similar to the wavelength
composition of the white prime disk than the wavelength of the blue desaturated screen phosphors. Perceptual-level priming, on the
other hand, targets the phenomenal colour properties of the stimuli. Phenomenally, desaturated blue annulus is more similar to the
white disk than desaturated green annulus. Our opponent may, then, argue that what is registered unconsciously by the visual system
is not the phenomenal colour (blue, green), but rather just the wavelength properties of the reflecting surfaces of the stimuli. The
results and reasoning of Breitmeyer et al. leads Lamme (2015) to conclude that the transition from wavelength analysis to the
perception of colour marks the transition from unconscious neural processing to conscious vision. If this account is correct, the dual
model of phenomenal consciousness will not work for phenomenal colours – a rather significant loss.
If this objection to the dual model of consciousness is correct, we would need to distinguish between those phenomenal qualities
that can occur unconsciously (such as shapes) and those that cannot (such as colours). But it would be premature to conclude that
there cannot be unconscious colours. In fact, the issue is far from settled. In contrast to the Breitmeyer experiments, an experiment of
Boyer, Harrison, and Ro (2005) revealed that in a temporary TMS-induced blindsight in healthy subjects, discrimination of colour
identity (not just of wavelength) is possible at above chance levels without awareness of the colour. Furthermore, Ro et al. (2009)
similarly report that colour, not just wavelength information, is processed in the absence of conscious awareness. In a metacontrast
priming experiment, they found that unconsciously seen colour of a disk speeds up response times of colour discrimination of the
congruently coloured masking annulus. Railo, Salminen-Vaparanta, Henriksson, Revonsuo, and Koivisto (2012) performed a TMS
study in which forced-choice unconscious discrimination of colours was above chance when participants reported consciously per-
ceiving some stimulus features, but not its color. Finally, Norman, Akins, Heywood, and Kentridge (2014), using a metacontrast
masking paradigm similar to Breitmeyer et al. (2004), demonstrated successful priming by unseen stimuli when the prime and the
mask shared surface colour, but they didn’t find a corresponding priming effect in terms of reflected colour (i.e., in terms of wa-
velength properties). All this evidence points to the existence of unconscious visual processing of colours.
Note also that the timing of the stimulus presentation in the experiments is of paramount importance. The time frame of the white
disk appearance in the Breitmeyer et al. experiments (14.3 ms) is simply too small for an unconscious phenomenal colour to be fully
formed: it takes between 60 and 100 ms for a signal to travel to V4 and other areas in the inferotemporal cortex where colours are

7
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

processed (Kravitz, Saleem, Baker, Ungerleider, & Mishkin, 2013). Given this, it is not surprising that the priming with white disks in
the Breitmeyer et al. design cannot utilize the phenomenal aspects of colour. On the other hand, if the time for processing colour is
sufficiently long, such as in the Boyer et al. experiment (86–114 ms), or in continuous flash suppression (up to many seconds), we can
presumably get fully fledged but unconscious phenomenal colours. If this line of thinking is sound, the thesis of unconscious phe-
nomenality is defensible even for colours, not just for shapes.
In summary, we are prepared to revise the thesis of unconscious phenomenality (“All phenomenal qualities can occur without
consciousness”) to allow for phenomenal qualities that cannot occur without consciousness, providing empirical evidence suggests
this is desirable. The evidence available so far points rather towards the version of the dual model of phenomenal consciousness
according to which all phenomenal qualities can occur without consciousness.

3.2. Cognitive neuroscience

On unitary models of phenomenal consciousness, both the formation of phenomenal character and the awareness of it are secured
by a single neural process. As we observed, this picture is based entirely on stubborn intuitions: nothing in the empirical data of
neurobiology, cognitive neuroscience or experimental psychology provides any support to the unitary approach. On the contrary, the
available evidence favours the dual approach. In particular, the neuroscientific evidence points to the need of separating the neural
mechanisms creating the phenomenal qualities and the mechanisms that make these qualities conscious. The argument of the pre-
ceding section can thus be complemented by this further evidence.
The dual model neatly dovetails with one of the most discussed theories of conscious visual perception in cognitive neuroscience,
the one put forward by Victor Lamme. The theory enables the separation of brain mechanisms creating the phenomenal character of a
mental state from the mechanisms making it conscious by decomposing the formation of a conscious visual state into two parts. The
first is a process of fast feedforward sweep, a one-way activation of all the areas in the visual system of the brain, which usually takes
less than 80 ms (although the areas are not reached strictly chronologically according to their place in functional hierarchy; e.g., the
area MT is reached around 15 ms before the area V2). The second process consists of recurrent neural activity. During this phase of
perception, higher visual centres send the signal back to the lower ones. This feedback activity occurs about 100 ms after stimulus
onset. These recurrent loops, supported by strong synchronous activity in the theta, beta, and gamma bands, make the visual percept
conscious (Lamme, 2006, 2010, 2015).
The pivotal argument for linking consciousness and recurrent neural activity is that disrupting top-down feedback input to V1
from higher regions impairs visual awareness (see Boehler, Schoenfeld, Heinze, & Hopf, 2008; Pascual-Leone & Walsh, 2001). Im-
portantly, even without recurrent neural activity the visual information may be processed fairly comprehensively: for example, the
neural activation typical for human face perception in the fusiform area can be observed in the fMRI scanner even when the per-
ception is not conscious. We propose that the fast feedforward sweep could be the neural mechanism producing unconscious visual
phenomenal qualities. The subsequent recurrent activity in the visual system then makes these qualities conscious.
Lamme’s theory thus provides the neural underpinning for the dual model of phenomenal consciousness. Lamme himself would
not agree with our interpretation of his theory as a specimen of the dual model of phenomenal consciousness. In Lamme (2015) he
subscribes to the orthodox view that conscious perceptual processes have phenomenal character whereas unconscious ones don’t.
Despite this fact, we believe that his account is more compatible with the dual approach. Feedforward activation of the structures in
the visual system sets up the phenomenal character of perceptual states, while recurrent neural loops make this phenomenal char-
acter conscious. These two processes are obviously different in regards to both the neural mechanisms involved and as to their time
profiles – the production of phenomenal qualities precedes their entry into consciousness by tens of milliseconds, which is a very
robust distinction from the neural point of view.
Now, Lamme’s theory, which we use to support the dual model of phenomenal consciousness, has been criticized. For example,
Tsuchiya and van Boxtel (2010) suggest that recurrent processing is not necessary for consciousness. They argue (for other criticisms
see the commentaries in Lamme, 2010) that gist perception of an unfamiliar and unexpected visual scene seems to be possible before
recurrence could even start, i.e., in a purely feedforward manner (see also ffytche & Zeki, 2011). By the same token, they hold that
recurrent processing seems to be insufficient for consciousness, as experiments on perceptual suppression indicate. These are, of
course, important criticisms. If they prove correct, we can continue defending the view that unconscious phenomenal qualities are
generated by feedforward neural processes while letting the nature of neural processes making these qualities conscious be elucidated
by further empirical research. In other words, the dual model of phenomenal consciousness itself is independent of the particulars of
Lamme’s theory of conscious vision. We are using this theory as a vivid illustration of a neuroscientific account, which is both well
supported by currently available evidence, and at the same time fully compatible with the dual model of consciousness.
However, there are other models in cognitive neuroscience sustaining the same idea. The popular Global Neuronal Workspace theory
of consciousness (Dehaene, 2014) is a case in point. The theory explains conscious perception as an entry of sensory information into a
network of brain nodes and regions (in parietal, temporal and prefrontal cortex) via long axonal projections (Dehaene & Changeux, 2011;
Dehaene & Naccache, 2001). Although this is (probably) not how its authors conceive of it, the sensory information might be of a phe-
nomenal nature, not just a non-phenomenal, non-qualitative signal. This unconscious phenomenal information is then made conscious by
its entry into the Global Workspace (GW). The distinction between phenomenal-character-producing and consciousness-conferring neural
processes, typical for the dual model of phenomenal consciousness, is preserved in this interpretation of the GW theory.
Suppose now that we interpret the GW theory in the standard unitary fashion. The ascent to the global workspace simultaneously
makes a mental state conscious and creates its phenomenal character. The problem with this picture is that the qualitative trans-
formation of content entering the global workspace is left unexplained by the theory. The core of the GW theory of consciousness is

8
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

simply that a content is made available for report and action by entering the global workspace and that this is what makes the content
conscious. Notice how the problem of explaining the qualitative transformation disappears when we take the GW theory to be an
instance of the dual model of consciousness. Phenomenal character of a state is generated in dedicated brain areas (visual, auditory,
etc.), and the ascent to the global workspace is the mechanism that makes all perceptual contents conscious.
Other hypotheses that seem to be coherent with the dual model include those based on the distinction between specific and non-
specific brain systems. Roughly, the specific systems are responsible for establishing the content to be experienced, while the non-
specific systems allow this content to become conscious by modifying the content-related activity. An early example of this approach
is Llinás, Ribary, Contreras, and Pedroarena (1998). The authors postulate that a conscious mental state is the result of coordinated
activation of specific and non-specific thalamocortical circuits. A more recent theory develops a similar idea, focusing especially on
the integration of content-determining mechanisms and mechanisms regulating the level of consciousness. Bachmann and Hudetz
(2014) suggest that consciousness in general, and perceptual conscious state in particular emerges as a result of interaction between
the following two systems. The first (called D) is a bottom-up, specific content-determining data system which gets the data from the
senses and provides the perceived content. The second (called L) is a non-specific neuromodulatory system sending intracortical top-
down associative signals by which the level of consciousness is regulated.10 Supporting this line of thinking, recent neurophysio-
logical evidence highlights the capacity of pyramidal neurons, in layer 5 of the neocortex, to integrate two kinds of inputs. With
considerable simplification we can say that the first kind of input consists of feedforward signals to the basal dendrites of pyramidal
cells and its function is to fix the perceptual content. The second, modulatory mechanism then influences whether this content will be
experienced consciously. This latter mechanism is called apical amplification, for it is implemented in the apical synapses of mature
pyramidal cells (Phillips, Larkum, Harley, & Silverstein, 2016).11 All this new neurobiological evidence may provide further support
to the dual model of consciousness because it enables to disentangle the processes responsible for preparing the perceptual content
(i.e., the phenomenal character) from those regulating the conscious experience of this content.
This subchapter offers the basics of a neural explanation of consciousness. But, one might ask, what is that a theory of? We accept
the view that consciousness is always consciousness of this or that set of phenomenal qualities and we disavow the possibility of “pure
consciousness”, postulated by some mystics and meditators, completely and utterly devoid of all phenomenal aspects. In other words,
a theory of consciousness, by our lights, is a theory of how the independently constituted phenomenal qualities become conscious for
the subject. So, even on the dual model, consciousness is tied with phenomenal qualities. Consciousness has to bite its teeth into
something, so to speak, and this will be some phenomenal qualities. However, and this is crucial, we want to block the inference that
the phenomenal qualities are so inextricably bound with consciousness that they cannot exist without it. By not making this in-
ference, we enable the divorce of consciousness and phenomenality even on the standard Nagelian picture of consciousness as what-
it’s-like-ness. What-it’s-like-ness only occurs at the conscious level, for to be in a completely unconscious mental state does not feel like
anything. But phenomenality, we argue, is broader than what-it’s-like-ness, however deeply entrenched may be the identification of
the two in the current philosophy of mind. To sum up, we can speak of the one-way dependence of consciousness and phenomenal
qualities: consciousness depends on qualities for its existence, whereas qualities can exist independently of consciousness.

4. Conclusions

The consensus, quite rare in philosophy and neuroscience, is that there are no phenomenal qualities outside of consciousness, and
moreover that phenomenality equals experiential what-it’s-like-ness. Our view is that a mental state may be complete in terms of its
neurophysiological as well as phenomenal properties, not just in its conscious form but also in its unconscious form. The dual model
of phenomenal consciousness, which we advocated in this paper, thus requires that the mechanisms producing phenomenal qualities
and the mechanisms making these qualities conscious be treated as separate.
We have argued for the dual model in light of four distinct arguments. The conceptual argument claims that the notion of an
unconscious phenomenal quality is plausible, despite some stubborn intuitions. The conceptual argument paves the way for the
methodological argument, which consists in the claim that empirical research must treat consciousness of a mental state as a variable.
The presence or absence of consciousness is then the sole distinguishing criterion of otherwise identical states. The neuropsycho-
logical argument adds to this the idea that observed reactions of patients with neuropsychological disorders are guided by already
formed but not consciously registered phenomenal qualities. It is an inference to the best explanation: the existence of unconscious
phenomenal states best accounts for the behaviour of subjects observed in psychological experiments. In neglect and blindsight, only
the mechanisms of consciousness itself are defunct. The mechanisms for the production of phenomenal character might still be intact.
Finally, the argument from cognitive neuroscience draws attention to the fact that the dual model of phenomenal consciousness is
fully compatible with some of the most prominent theories of consciousness put forward by cognitive neuroscientists. These theories
provide a possible neural underpinning for the dual model, making it more empirically respectable.12

10
Bachmann and Hudetz underline that their thesis is in accord with many psychophysical experiments involving visual masking, metacontrast masking, perceptual
latency priming, motion-induced blindness or Ganzfeld effect.
11
Remarkably, much of the input to apical synapses is provided by top-down, feedback or recurrent connections that Lamme’s theory (briefly described above in
this section) also sees as playing a crucial role in perceptual states’ becoming conscious.
12
The dual model is also fully compatible with the hotly debated paradigm of contemporary cognitive science, the Predictive Coding (or Predictive Processing)
framework (Clark, 2013; Cleeremans, 2014; Hohwy, 2013; Melloni, 2015). On this account, probabilistic perceptual hypotheses are formed unconsciously and are then
pushed into consciousness by a different mechanism (so far left unspecified by the proponents of the framework). Panichello, Cheung, and Bar (2013) stress the
compatibility of the Predictive Coding framework and Lamme’s theory of conscious perception.

9
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

Together, these arguments make a very strong case for the dual model of consciousness. Compared to that, the arguments for the
currently dominant unitary model seem to be less persuasive. For instance, how are the discriminatory reactions of neuropsycho-
logical patients to be explained by the unitary model? Defenders of the unitary model could say that neuronal activity completely
devoid of phenomenality may still account for the above-noted behaviours of neglect and blindsight subjects. When the subject
indicates, e.g., that she would not live in the house consumed by flames, she does not harbour any unconscious phenomenal snapshot
of the flames. The behaviour is completely automated, reflex-like; it’s a neural push and pull between visual and speech centres.
Unconscious perception is nothing over and above transduction of information via sensory channels in the brain. No phenomenality,
just neurons firing in the dark. Although we cannot prove beyond all doubt that the discriminations could not be performed without
unconsciously registered phenomenal qualities, we believe that the just-neurons-firing-in-the-dark story is not very appealing. It is
unclear whether it really explains anything. In contrast, the dual model provides a parsimonious explanation of how unconscious
discrimination and recognition is performed: it is performed by means of the same phenomenal perceptual qualities as appear at the
conscious level. The work in cognitive neuroscience supporting our dual approach (Section 3.2) then enhances plausibility of the dual
model by suggesting that the phenomenal character of a conscious mental state is identical to the phenomenal character of its
unconscious counterpart. These phenomenal characters are the same in its conscious and unconscious form because they are created by
the same neural mechanism.13
Similarly, the methodological desideratum that consciousness is to be treated as an experimental variable is also not satisfied by
the unitary model of phenomenal consciousness. According to this model, conscious mental states possess phenomenality while
unconscious mental states lack it completely. The two types of states thus differ more substantively than is permitted by the desi-
deratum. The dual model, on the other hand, does justice to this maxim of empirical research, for it claims that conscious and
unconscious states differ only in the presence of consciousness. It could be objected, though, that the dual model doesn’t satisfy
another methodological principle, that of double dissociation. The dual model allows phenomenal qualities without consciousness,
but is silent about the reverse possibility of consciousness without phenomenal qualities.14 This objection is valid. As we already
stated above, consciousness must be consciousness of something, of this or that phenomenal quality. Double dissociation of con-
sciousness and phenomenal quality is therefore impossible, for there cannot be consciousness without phenomenal qualities. In this
respect, then, our model of consciousness is methodologically sub-optimal. However, the requirement that the phenomenon under
study be treated as a variable is scientifically more important, and the dual proposal accommodates it fully.
The disadvantage of the unitary model of phenomenal consciousness is that it makes phenomenal consciousness intractable. What
truly magical brain process can produce a phenomenal quality and turn the spotlight of consciousness upon it at the same time?
Distinguishing the mechanisms for phenomenal qualities and for consciousness along the lines of the dual model looks like a more
feasible project than the attempt to explain them en bloc – an attempt those who prefer the unitary model are bound to undertake.
One can also think about a hybrid unitarian theory, according to which we can theoretically separate consciousness from perceptual
qualities, but in reality they always co-occur. We see three disadvantages of this view. First, one needs to secure that both processes –
consciousness and the production of qualities – always and in all circumstances co-occur together. But it is unclear how to secure this.
Second, we cannot find any empirical support for the claim of co-occurrence of consciousness and perceptual quality. The dual model
we favour, in contrast, has some support in current cognitive neuroscience. Lamme’s theory of neural recurrence does separate the
two processes in time and space. Third, it is not clear what mechanisms can account for unconscious perception (provided one accepts
its existence) on the hybrid view. If the phenomenal qualities-mechanism can only operate when consciousness is present, un-
conscious perception must recruit a different set of mechanisms than those enabling conscious perception (which trades in phe-
nomenal information). What are these mechanisms and why are they not recruited in cases of conscious perception? The hybrid
proposal leaves this unexplained. The dual model of phenomenal consciousness is free of these perplexities. But it is not just an
attractive theoretical option. It suggests a way forward for the empirical research on consciousness as well. This research could be
split into the search for the mechanisms producing phenomenal character of a mental state, and an inquiry into what makes such state
conscious.15

References

Arend, I., Rafal, R., & Ward, R. (2008). Spatial and temporal deficits are regionally dissociable in patients with pulvinar lesions. Brain, 131, 2140–2152.
Baars, B. (1997). In the theatre of consciousness. Journal of Consciousness Studies, 4, 292–309.
Baars, B. (2003). Introduction: Treating consciousness as a variable: The fading taboo. In B. J. Baars, W. P. Banks, & J. B. Newman (Eds.). Essential sources in the
scientific study of consciousness (pp. 1–10). Cambridge, MA: MIT Press.
Bachmann, T., & Hudetz, A. G. (2014). It is time to combine the two main traditions in the research on the neural correlates of consciousness: C = L x D. Frontiers in
Psychology, 5, 1–13.
Bain, A. (1894). Definition and problems of consciousness. Mind, 3, 348–361.
Berti, A. (2002). Unconscious processing in neglect. In H.-O. Karnath, D. A. Milner, & G. Vallar (Eds.). The cognitive and neural bases of spatial neglect (pp. 313–326).
Oxford: Oxford University Press.
Berti, A., & Rizzolatti, G. (1992). Visual processing without awareness: Evidence from unilateral neglect. Journal of Cognitive Neuroscience, 4, 345–351.
Bisiach, E., & Rusconi, M. L. (1990). Break-down of perceptual awareness in unilateral neglect. Cortex, 26, 643–649.
Block, N. (2001). Paradox and cross purposes in recent work on consciousness. Cognition, 79, 197–219.

13
In Lamme’s theory, this would be the fast feedforward sweep.
14
We thank Juraj Hvorecký for pressing this objection in conversation.
15
We thank participants at the Ernst Mach Workshop V and to various audiences at conferences and workshops in the Czech Republic for their helpful criticisms of
the previous versions of this paper. Our special thanks for continuous discussion and detailed comments on our ideas go to Juraj Hvorecký.

10
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

Block, N. (2011). The higher order approach to consciousness is defunct. Analysis, 71, 419–431.
Block, N., & Phillips, I. (2016). Debate on unconscious perception. In B. Nanay (Ed.). Current controversies in philosophy of perception (pp. 165–192). New York:
Routledge.
Boehler, C. N., Schoenfeld, M. A., Heinze, H. J., & Hopf, J. M. (2008). Rapid recurrent processing gates awareness in primary visual cortex. Proceedings of the National
Academy of Sciences of the United States of America, 105, 8742–8747.
Boyer, J. L., Harrison, S., & Ro, T. (2005). Unconscious processing of orientation and colour without primary visual cortex. Proceedings of the National Academy of
Sciences of the United States of America, 102, 16875–16879.
Breitmeyer, B. G. (2015). Psychophysical “blinding” methods reveal a functional hierarchy of unconscious visual processing. Consciousness and Cognition, 35, 234–250.
Breitmeyer, B. G., Ro, T., & Singhal, N. S. (2004). Unconscious color priming occurs at stimulus—not percept-dependent levels of processing. Psychological Science, 15,
198–202.
Brogaard, B. (2011). Are there unconscious perceptual processes? Consciousness and Cognition, 20, 449–463.
Byrne, A. (1997). Some like it HOT: Consciousness and higher-order thoughts. Philosophical Studies, 2, 103–129.
Campion, J., Latto, R., & Smith, Y. M. (1983). Is blindsight an effect of scattered light, spared cortex, and near threshold vision? Behavioral and Brain Sciences, 6,
423–486.
Carruthers, P. (2010). Introspection: Divided and partly eliminated. Philosophy and Phenomenological Research, 80, 76–111.
Clark, A. (2000). A theory of sentience. Oxford: Oxford University Press.
Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36, 181–204.
Cleeremans, A. (2014). Prediction as a computational correlate of consciousness. International Journal of Anticipatory Computing Systems, 29, 3–13.
Coleman, S. (unpublished results). Unconscious qualities as the basis of content.
Cowey, A. (2010). The blindsight saga. Experimental Brain Research, 200, 3–24.
Crick, F., & Koch, Ch. (2003). A framework for consciousness. Nature Neuroscience, 6, 119–126.
Damasio, A. (1999). The feeling of what happens: Body and emotion in the making of consciousness. New York: Harcourt Brace.
Dehaene, S. (2014). Consciousness and the brain: Deciphering how the brain codes our thoughts. New York: Penguin Books.
Dehaene, S., & Changeux, J.-P. (2011). Experimental and theoretical approaches to conscious processing. Neuron, 70(2), 200–227.
Dehaene, S., & Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: Basic evidence and a workspace framework. Cognition, 79, 1–37.
Dretske, F. (2006). The epistemology of pain. In M. Aydede (Ed.). Pain: New papers on its nature and the methodology of its study (pp. 59–73). Cambridge, MA: MIT Press.
ffytche, D. (2000). Imaging conscious vision. In T. Metzinger (Ed.). Neural correlates of consciousness (pp. 221–230). Cambridge, MA: MIT Press.
ffytche, D. H., & Zeki, S. (2011). The primary visual cortex, and feedback to it, are not necessary for conscious vision. Brain, 134, 247–257.
Gazzaniga, M. S., Ivry, R. B., & Mangun, G. R. (2014). Cognitive neuroscience: The biology of mind. New York: W. W. Norton & Company.
Goff, P. (2016). The phenomenal bonding solution to the combination problem. In G. Brüntrup, & L. Jaskolla (Eds.). Panpsychism: Contemporary perspectives (pp. 283–
303). Oxford: Oxford University Press.
Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15, 20–25.
He, S., Cavanagh, P., & Intriligator, J. (1996). Attentional resolution and the locus of visual awareness. Nature, 383, 334–337.
Hohwy, J. (2013). The predictive mind. Oxford: Oxford University Press.
Husain, M., & Rorden, Ch. (2003). Non-spatially lateralized mechanisms in hemispatial neglect. Nature, 4, 26–36.
Kihlstrom, J. F. (1987). The cognitive unconscious. Science, 237, 1445–1452.
Koch, Ch. (2004). The quest for consciousness: A neurobiological approach. Englewood: Roberts & Company Publishers.
Kozuch, B. (2014). Prefrontal lesion evidence against higher-order theories of consciousness. Philosophical Studies, 167, 721–746.
Kravitz, D. J., Saleem, K. S., Baker, C. I., Ungerleider, L. G., & Mishkin, M. (2013). The ventral visual pathway: An expanded neural framework for the processing of
object quality. Trends in Cognitive Sciences, 17, 26–49.
Kripke, S. (1980). Naming and necessity. Oxford: Basil Blackwell.
Lamme, V. (2006). Towards a true neural stance on consciousness. Trends in Cognitive Sciences, 10, 494–501.
Lamme, V. (2010). How neuroscience will change our view on consciousness. Cognitive Neuroscience, 1, 204–240.
Lamme, V. (2015). The crack of dawn – Perceptual function and neural mechanisms that mark the transition from unconscious processing to conscious vision. In T.
Metzinger, & J. M. Windt (Eds.). Open MIND (pp. 1–34). Frankfurt am Main: MIND Group.
Llinás, R., Ribary, U., Contreras, D., & Pedroarena, C. (1998). The neuronal basis for consciousness. Philosophical Transactions of the Royal Society of London, B,
353(1377), 1841–1849.
Marshall, J. C., & Halligan, P. W. (1988). Blindsight and insight in visuo-spatial neglect. Nature, 336, 766–767.
Mazzi, C., Bagattini, C., & Savazzi, S. (2016). Blind-sight vs. degraded-sight: Different measures tell a different story. Frontiers in Psychology, 7, 901.
Melloni, L. (2015). Consciousness as inference in time. A commentary on Victor Lamme. In T. Metzinger, & J. M. Windt (Eds.). Open MIND (pp. 1–13). Frankfurt am
Main: MIND Group.
Merikle, P. M., Smilek, D., & Eastwood, J. D. (2000). Perception without awareness: Perspective from cognitive psychology. Cognition, 79, 115–134.
Moutoussis, K., & Zeki, S. (2002). The relationship between cortical activation and perception investigated with invisible stimuli. Proceedings of the National Academy of
Sciences of the United States of America, 99, 9527–9532.
Neuman, Y. (2004). What does pain signify? A hypothesis concerning pain, the immune system and unconscious pain experience under general anesthesia. Medical
Hypotheses, 63, 1051–1053.
Norman, L. J., Akins, K., Heywood, C. A., & Kentridge, R. W. (2014). Color constancy for an unseen surface. Current Biology, 24, 2822–2826.
Overgaard, M. (2012). Blindsight: Recent and historical controversies on the blindness of blindsight. WIREs Cognitive Science, 3, 607–614.
Overgaard, M., Fehl, K., Mouridsen, K., & Cleeremans, A. (2008). Seeing without seeing? Degraded conscious vision in a blindsight patient. PLoS ONE, 3, 1–4.
Overgaard, M., & Grünbaum, T. (2011). Consciousness and modality: On the possible preserved visual consciousness in blindsight subjects. Consciousness and Cognition,
20, 1855–1859.
Panichello, M. F., Cheung, O. S., & Bar, M. (2013). Predictive feedback and conscious visual experience. Frontiers in Psychology, 3, 1–8.
Pascual-Leone, A., & Walsh, V. (2001). Fast backprojections from the motion to the primary visual area necessary for visual awareness. Science, 292, 510–512.
Peters M. A. K., & Lau, H. (2015). Human observers have optimal introspective access to perceptual processes even for visually masked stimuli. eLIFE, 4, e09651, 1–30.
Phillips, I. (2016). Consciousness and criterion: On Block’s case for unconscious seeing. Philosophy and Phenomenological Research, 93, 419–451.
Phillips, W. A., Larkum, M. E., Harley, C. W., & Silverstein, S. M. (2016). The effects of arousal on apical amplification and conscious state. Neuroscience of
Consciousness, 1, 1–13.
Pöppel, E., Held, R., & Frost, D. (1973). Residual visual function after brain wounds involving the central visual pathways in man. Nature, 243, 295–296.
Prinz, J. (2012). The conscious brain. New York: Oxford University Press.
Prinz, J. (2015). Unconscious perception. In M. Matthen (Ed.). The Oxford handbook of philosophy of perception (pp. 371–393). Oxford: Oxford University Press.
Prinz, J. (2017). Unconscious vision and the function of consciousness. In Z. Radman (Ed.). Before consciousness: In search of the fundamentals of mind (pp. 142–162).
Exeter: Imprint Academic.
Railo, H., Andersson, E., Kaasinen, V., Laine, T., & Koivisto, M. (2014). Unlike in clinical blindsight patients, unconscious processing of chromatic information depends
on early visual cortex in healthy humans. Brain Stimulation, 7, 415–420.
Railo, H., Salminen-Vaparanta, N., Henriksson, L., Revonsuo, A., & Koivisto, M. (2012). Unconscious and conscious processing of color rely on activity in early visual
cortex: A TMS study. Journal of Cognitive Neuroscience, 24, 819–829.
Revonsuo, A. (2010). Consciousness: The science of subjectivity. Hove: Psychology Press.
Ro, T., Singhal, N. S., Breitmeyer, B. G., & Garcia, J. O. (2009). Unconscious processing of color and form in metacontrast masking. Attention,
Perception, & Psychophysics, 71, 95–103.
Rosenthal, D. M. (1986). Two concepts of consciousness. Philosophical Studies, 49, 329–359.

11
T. Marvan, M. Polák Consciousness and Cognition 56 (2017) 1–12

Rosenthal, D. M. (2005a). Consciousness and mind. Oxford: Clarendon Press.

Rosenthal, D. M. (2010). How to think about mental qualities. Philosophical Issues, 20, 368–393.
Rosenthal, D. M. (2011). Exaggerated reports: Reply to Block. Analysis, 71, 431–437.
Rosenthal, D. M. (2012). Higher-order awareness, misrepresentation and function. Philosophical Transactions of the Royal Society B, 367, 1424–1438.
Rosenthal, D. M. (1997). A theory of consciousness. In N. Block, O. Flanagan, & G. Güzeldere (Eds.). The nature of consciousness: Philosophical debates (pp. 729–753).
Cambridge, MA: MIT Press.
Rosenthal, D. M. (2005b). Sensory qualities, consciousness, and perception. In D. M. Rosenthal (Ed.). Consciousness and mind (pp. 175–226). Oxford: Clarendon Press.
Sahraie, A., Hibbard, P. B., Trevethan, C. T., Ritchie, K. L., & Wieskrantz, L. (2010). Consciousness of the first order in blindsight. Proceedings of the National Academy of
Sciences of the United States of America, 107, 21217–21222.
Sandberg, K., & Overgaard, M. (2015). Using the perceptual awareness scale (PAS). In M. Overgaard (Ed.). Behavioral methods in consciousness research (pp. 181–195).
Oxford: Oxford University Press.
Sandberg, K., Timmermans, B., Overgaard, M., & Cleeremans, A. (2010). Measuring consciousness: Is one measure better than the other? Consciousness and Cognition,
19, 1069–1078.
Tamietto, M., Cauda, F., Corazzini, L. L., Savazzi, S., Marzi, C. A., Goebel, R., ... de Gelder, B. (2010). Collicular vision guides nonconscious behavior. Journal of
Cognitive Neuroscience, 22, 888–902.
Tsuchiya, N., & van Boxtel, J. J. (2010). Is recurrent processing necessary and/or sufficient for consciousness? Cognitive Neuroscience, 1, 230–231.
Vallar, G., & Perani, D. (1986). The anatomy of unilateral neglect after right-hemisphere stroke lesions. A clinical/CT-scan correlation study in man. Neuropsychologia,
24, 609–622.
Vallar, G., Rusconi, M. L., & Bisiach, E. (1994). Awareness of contralesional information in unilateral neglect: Effects of verbal cueing, tracking and vestibular
stimulation. In M. Moscovitch, & C. A. Umiltà (Eds.). Attention and performance XV: Conscious and nonconscious information processing (pp. 377–391). Cambridge,
MA: MIT Press.
Weiskrantz, L. (1986). Blindsight: A case study and implications. Oxford: Oxford University Press.
Weiskrantz, L. (1996). Blindsight revisited. Current Opinion in Neurobiology, 6, 215–220.
Weiskrantz, L. (2009). Is blindsight just degraded normal vision? Experimental Brain Research, 192, 413–416.
Young, B., Keller, A., & Rosenthal, D. M. (2014). Quality-space theory in olfaction. Frontiers in Psychology, 5, 1.

12