When the aim is to provide an account of the difference between a representation’s being phenomenally conscious and it’s being unconscious many philosophers are drawn to some form of representationalism. This is motivated in part by the prospect of reducing the problem of consciousness to the problem of intentionality or representation (Tye 1995; Dretske 1995; Rosenthal 2005; Lycan 1996; Metzinger 2003; Kriegel 2009; Kriegel & Williford 2006). But many of those who are dissatisfied with a representational criterion argue that the difference is due to some sort of integration (Dehaene 2014; Van Gulick 2004; Edelman & Tononi 2000; Damasio 2010; Metzinger 1995; Kant 1999; Schlicht 2011). Such integration may eventually result in a higher-order or more complex representational state. In that sense, the two accounts do not mutually exclude each other. But they give different answers to the question of what is responsible for the representation exhibiting the feature of being conscious. To put forward both a representational condition and an integration mechanism would amount to wearing a belt as well as suspenders. Williford’s paper demonstrates that other theories are also possible. He favors self-reflexivity as the core feature a representation must exhibit in order for it to be conscious.
In the first part of his paper, Williford scrutinizes all dominant varieties of representationalism, especially with respect to their explanatory power regarding the subjective character of conscious experiences. His case against first-order, higher-order, and same-order or self-representationalism is solid, and I have nothing to add in this regard (cf. also Schlicht 2008b; Vosgerau et al. 2008).[5]
The basis for answering the question as to which conditions have to be met by a single sensation of red in order for it to be conscious and subjectively experienced is the observation that the organism in question is already conscious in the creature-sense. This general consciousness (or state of vigilance) admits of degrees (from deep coma to wakefulness) and is one of the conditions for being able to enjoy a sensation of red at all (Dehaene et al. 2006). Empirical evidence points to the assumption that the neural structures in the brain supporting this state contain the relevant structures monitoring and regulating the homeostatic balance of the whole organism. Damasio (1999, 2010) calls these structures “proto-self”-structures, the biological forerunner of that which we eventually experience as a sense of self. He assumes that the brain can only perform these functions of monitoring and regulating if the overall state of the whole organism is represented in the brain.
In addition to representations of the organism, the brain is assumed to produce representations of (objects in) the external world. Given the limited capacity of conscious perception and memory systems, such representations stand in competition (Koch 2004). The basic idea of integration-theories is that some of these competing representations, like a sensation of red, are conscious because they are integrated into a more global state that also contains the structures responsible for creature-consciousness. Van Gulick (2004) has sketched such an integration-theory, based on ideas already to be found in Metzinger (1995):
The basic idea is that lower-order object states become conscious by being incorporated as components into the higher-order global states (HOGS) that are the neural and functional substrates of conscious self-awareness. The transformation from unconscious to conscious state is not a matter of merely directing a separate and distinct meta-state onto the lower-order state but of “recruiting” it into the globally integrated state that is the momentary realization of the agent’s shifting transient conscious awareness. (Van Gulick 2004, pp. 76-77)
In other words, a single sensation of red is consciously experienced if the neural activation pattern supporting this sensation is integrated in the right way into the neural basis representing the overall state of the organism, the “dynamic core” in Edelman’s words (Edelman & Tononi 2000).[6]
Importantly, the integration mechanism (which is what has to be determined empirically in this framework)—synchronous oscillations, say— is not only responsible for producing a coherent single experiential state of the organism; it also thereby conveys subjective character to the integrated individual representations. If this idea is combined with Damasio’s (1999) notion of proto-self-structures, then integration facilitates a strong connection between the substrate of an individual sensation (of red, say) and the biological structure representing the organism in the brain.[7] Of course, just like on all other theories, the hard problem is not addressed head-on, i.e., it is not explained why activation of these structures feels like something at all. All that can be provided (at this stage anyway) is a coherent story of how all these aspects hang together. But one advantage of the present integration-account is that by establishing a connection between the organism (as represented in the brain) and its object-representations we can make sense of the important fact that all conscious representations feel like something for the organism. The organism provides, as Damasio puts it, a “haven of stability and invariance” (1999, p. 142, p. 153; see also Metzinger 2003, p. 161), i.e., just what we need in order to account for subjective character. For remember that subjective character is the feature that remains stable across different representations, while qualitative character is the feature that distinguishes different representations from each other. So in order to get an account of subjective character started, we have to look for the point of “maximal invariance of content in the conscious model of reality”, as Metzinger (2003, p. 134) puts it. Metzinger agrees that this invariance is most likely due to the organism and its bodily structures represented in the brain, since it is invariance (or maintenance of homeostatic balance) that keeps the organism alive. Another advantage of this view is that it does so without introducing a questionable new entity and by avoiding Williford’s phenomenologically counterintuitive claim that the stream of consciousness should be identified with the subject of experience. In this commentary, I cannot argue in detail for this positive alternative but hope that these sketchy comments give the reader a general idea of what it aims at. Since I am dissatisfied with Williford’s identification of the subject of experience with the stream or an episode of consciousness, let me now finally turn to an argument for a different conceptualization of the subject.