[1]
As will become clear shortly, contrary to ordinary ways of speaking, I do not hold that persons must be the “subject relata” of acquaintance relations. Rather, I hold that episodes of consciousness are, fundamentally, the subject relata.
[2]
This is not to imply that scientific realism entails physicalism, of course.
[3]
This is a difficult issue I will not enter into. See e.g., Dainton (2000, 2008); Strawson (2009).
[4]
I will occasionally use the terms “reflexivity” and “reflexive awareness” to denote just this characteristic of consciousness (i.e., that of its always being aware of itself). It is not to be confused with “reflection” in the sense of introspection. It is more like the logical usage of “reflexive” (as in “reflexive relation”). The acquaintance relation is reflexive on the domain of conscious states, according to the view accepted here (as well as being anti-symmetric). But not everything that stands in this relation is self-acquainted—episodes of consciousness are, but they are also acquainted with sensory qualities, and these latter are not acquainted with anything.
[5]
The Grain Problem, customarily attributed to Wilfred Sellars, is a problem for any identity theory according to which sensory qualities are really brain properties of some sort. Roughly put, the problem is that brain properties are complex and structured while sensory qualities seem, on the face of it, ultimately simple and unstructured. For good discussions with references to Sellars see Clark (1989) and Lockwood (1993).
[6]
See e.g., Gennaro (2012); Kriegel (2006, 2009); Weisberg (2008, 2014). Williford (2006) can be taken to express a pure S view—the conscious mental state has itself for its own object, not some portion of itself. We can also classify Carruthers as an S theorist; see Carruthers (2000, 2005). Gennaro would not describe himself as an S theorist. S theory is also often called self-representationalism.
[7]
For naturalistic versions, see e.g., Damasio (1999 and 2010), Metzinger (2004), and Sebastian (forthcoming). I am sure that Damasio, Metzinger, and Sebastian would reject this label, but the point of it is that all these theories identify subjective consciousness, in one way or another, with the representation of a “self,” understood in a naturalistically acceptable sense. See e.g., Metzinger (2004), p. 302: “In short, a self-model is a model of the very representational system that is currently activating it within itself” (emphasis original); and Damasio (2010), p. 180: “… [T]he brain constructs consciousness by generating a self process within an awake mind. The essence of the self is a focusing of the mind on the material organism that it inhabits.” It should be noted that Metzinger allows that there could be conscious experience that does not involve subjective character (see Metzinger 2004, pp. 559-560). Thus my categorization here applies at most only to his theory of subjective consciousness. Since, for me (as for Damasio), all consciousness necessarily has subjective character, this difference in detail will not loom large in what follows.
[8]
See the excellent discussion of the “non-conceptual content” literature in Hopp (2011).
[9]
For example, Tononi & Koch (2008, pp. 240–241) do not seem to think that the “sense of self” is essential (though Tononi (2014) may have recently changed his view); Damasio (1999, 2010) is in the opposing camp; see also Northoff (2013).
[10]
I’ll not go into the Higher-Order Thought vs. Higher-Order Perception debate. See e.g., Gennaro (2012).
[11]
See e.g., Neander (1998); Horgan & Kriegel (2007); Weisberg (2008); Tye (2011, pp. 4–8). See Kidd (ms) for an excellent discussion of these epistemological issues in the (not interestingly different) case of S theory.
[12]
See Mandik (2009 and forthcoming) on the “Unicorn problem” and Block (2011). See Rosenthal (2011, 2012); Weisberg (2011a, 2011b); Kiefer (2012); Wilberg (2010), and Berger (2013) for discussions of various strategies for dealing with Higher-Order-Thoughts (HOTs) without Lower-Order-Thoughts (LOTs).
[13]
What I am calling the “non-literal” interpretation is, in effect, the position in Berger (2013). And in Rosenthal (2011, p. 436) he in effect claims that the non-literal position (as I am calling it) has always been his view. See Mandik (forthcoming) on this.
[14]
I assume here but will not argue that teleosemantics is the most plausible naturalistic theory of content. There may be other naturalistic options that allow one to make good sense of the notion of essential indexicality in a way that could help H theory here, but I doubt it.
[15]
See Sebastian (forthcoming) for a Damasio-inspired turn toward a P theory (at least, that was my interpretation of it).
[16]
We can’t eliminate the misrepresentation problem, however. But we bracket that for now. See Kidd (ms) and Weisberg (2008).
[17]
A la Higginbotham (2003 and 2010) and before that (implicitly) Smullyan (1984); see Cappelen & Dever (2013, pp. 160-161). The hyperset model in Williford (2006) is the skeleton of such a theory. See also Kapitan (2006).
[18]
See Cappelen & Dever (2013, ch. 10). They attempt to capture this grain by appealing to relatively un-puzzling epistemic limitations. I believe they are on the right track, even if I would characterize the specific limitations in question a bit differently (see the discussion below on our ignorance of what fundamentally individuates us).
[19]
See Henrich (1982); Frank (2002, 2007); Shoemaker (1968). The issue, which is part of the “essential indexical” problematic, is, when put into a “self-model theory” context (which is not to be identified with Metzinger’s views), just that modeling something structurally isomorphic to oneself is not sufficient for knowledge that one is modeling oneself, as opposed to having behavioral control through such an interface (I could be controlling my doppelgänger unwittingly and just as effectively). One would need to know that the thing modeled is oneself (and not something else that happens to be isomorphic to it, like one’s counterpart in a close possible world). One cannot, on pain of regress, derive such knowledge from a set of descriptions of oneself without already knowing that at least one of the descriptions does indeed apply to oneself. So one must have some direct self-knowledge, such as knowledge by acquaintance that one is the relevant so-and-so. An S theory wedded to a teleosemantic theory of representation and externalist theory of justification has the advantage of being able to accommodate direct reference and non-inferential knowledge of oneself, though one will regard this as a mere simulacrum of the phenomenology.
[20]
See e.g., Tye (2000, ch. 6). I will not be able to go into the back and forth over Swampman. Suffice it to say that despite hearing many attempted rejoinders over the years, I still find the objection to be a reductio of representationalist theories of consciousness wedded to historico-externalist theories of content.
[21]
See e.g., Carruthers (2000, 2005) and Gennaro (2012, pp. 45–49). Briefly, the sort of functional role semantics Carruthers embraces derives actual, occurrent content from dispositions, and it is actually subject to variations on the Swampman theme.
[22]
See Kriegel (2007) for an excellent discussion of phenomenological impasses. Thanks to Jennifer Windt for reminding me of this lucid article.
[23]
For just a few examples of the historical pedigree here, see Caston (2002) on Aristotle, Williams (2000) and Coseru (2012, ch. 8) on the Indian and Buddhist debate, Thiel (2011) on the early modern problematic, Frank (2004) on the German Idealist and Romantic discussion, and Zahavi (1999 and 2006) on the Phenomenological movement.
[24]
See e.g., Lane & Liang (2011). (Thanks to an anonymous reviewer for pointing this nice article out to me.) If, as I shall argue, subjective character is not fundamentally a representational matter at all, the issue of representational immunity to error through misidentification is orthogonal. To the extent that the attribution of ownership is a representational matter, it may or may not be possible to misattribute ownership, as far as the view defended here is concerned.
[25]
The terminology apparently derives from Prufer (1975) and is very common in phenomenological quarters. See e.g., Zahavi (1999); Crowell (2011, p. 16).
[26]
See Moore (1910); Butchvarov (1979, p. 250, 1998, p. 55), and Williford (2004). On Hume in this regard, see Strawson (2011).
[27]
Cf. Moore (1910, p. 57). (This paper of Moore’s is not as well known as his “Refutation of Idealism,” but it deserves to be.)
[28]
I will not attempt to offer an account of the (synchronic or diachronic) unity of consciousness in this paper (again, see e.g., Dainton 2000) or of mereological principles governing “parts” of episodes of consciousness and episodes as “wholes.” It is enough for my purposes that one recognize that conscious episodes are internally variegated unities of some sort.
[29]
I defend this view also in Williford (2011a, 2011b) and in Williford et al. (2012); Dreyfus (2011) is an articulation and defense of a similar view from a Buddhist perspective.
[30]
See e.g., Kriegel (2011) and the papers in Kriegel (ed.) (2013), as well as Kriegel’s excellent introduction to that volume.
[31]
We could possibly hold that even if the property instances are not real, the universals represented are, and try to account for the difference in phenomenology in terms of those real differences. But this sort of view does not allow us to make sense of the concrete but hallucinatory representation of different particular instances of the different properties.
[32]
I have briefly made similar arguments in Williford (2013).
[33]
In particular, a representationalist could say that the represented difference between the pink and purple clouds is just as hallucinatory as the clouds themselves. This is, in a sense, correct. However, representationalists hold (or ought to hold, anyway) that phenomenal differences always correlate with differences in the representations themselves (and only normally in the objects of representation). If there are phenomenal differences, there exist some differences inherent in consciousness that are not merely the objects of representation. What I am claiming is that we are acquainted with this differentiation under the guise of differences in objects represented. An adherent of the Transparency Intuition would deny this, of course. And I don’t take these considerations to constitute a knock-down argument. (Thanks to an anonymous reviewer for bringing this up.)
[34]
I’ve argued this is in a bit more detail in Williford (2013). For relevant background ideas see Williford (2005 and 2007). For a discussion of the Revelation Thesis see e.g., Stoljar (2006, ch. 11) and Goff (forthcoming).
[35]
I have considered our acquaintance with a differentiated manifold qua mind-world boundary in more detail in Williford (2013).
[36]
I have briefly argued this before in Williford (2011b). I was pleased to find that a similar line of argument was pursued by the eleventh-century Buddhist philosopher Ratnakīrti; see Ganeri (2012, p. 217).
[37]
Previous episodes of consciousness normally connected to the present episodes (the ones producing this document) found themselves trying to live with the latter horn of the dilemma in the flawed Williford (2005).
[38]
This is demonstrable. First, obviously, there is no logical problem with reflexive relations. Second, it requires special and highly questionable premises to generate another regress here. See Williford (2006). See also Kriegel (2009, p. 124) and Janzen (2008, p. 110).
[39]
The knife blade and candle flame competing analogies loom large in the Indo-Tibetan debate on this issue. Clearly, the analogies will be found, by opponents and proponents, to be exactly as plausible as the views they encode.
[40]
Though even Rosenthal’s own view was pushed into being (or always was) problematic in this regard, as noted above.
[41]
Here I am in considerable agreement with Langsam (2011, ch. 3).
[42]
They are phenomenologically indistinguishable in the way that the stream of consciousness’s being temporally continuous is, plausibly, phenomenologically indistinguishable from consciousness’s being punctate or discrete, or in the way in which consciousness’s seeming free from causal determination is phenomenologically indistinguishable from its simply not seeming determined (because the causal relations are inaccessible, as Spinoza suggested).
[43]
See the following links:
https://www.youtube.com/watch?v=LXOqD5B5Sxc
http://www.deceptology.com/2010/10/headless-woman-illusion.html
[44]
While Hofstadter’s Gödel-inspired model might be problematic (both in terms of physical implementation and in terms of the strong mathematical realism it might presuppose), it is certainly in the right class of models we should be considering. See Hofstadter (2007).
[45]
For discussion, see Chalmers (1996) and Buechner (2008, ch. 3).