In the first part of this piece, I opined that self-consciousness consists of an entity’s individuality qua individuated by neurobiological structure, and illustrated by contrast its unique feature of being neither simultaneously nor disjunctively predicable to external objects. I defended this definition by appealing to its conceptual parsimony compared to other potential definitions of self-consciousness, while also dealing with the main contention made against this definition, namely the purported existence of irreducible qualia. To begin this second and final part, I will present a variety of positive arguments supporting this definition, drawing on the work of other philosophers on the subject, and, finally, explaining its relevance to that concept that lends itself to the title of this series— death.
That individuality can neither be simultaneously nor disjunctively predicated of objects finds its philosophical justification in the work of G.E.M Anscombe, particularly as interpreted by Sebastian Rödl— self-consciousness amounts to what Rödl calls a spontaneous material substance concept, one that distinguishes itself from the schema of receptive reference— the type of reference subject to the principle of the identity of discernables, as I outlined in part one. Rödl, in his aptly named 2007 monograph Self-consciousness, defines receptive reference as:
…a concept of reference according to which referring to an object is relating to it in a way that gives application to the notion of getting the object right. This is receptive reference, reference mediated by an act of receptivity. In the fundamental case, reference depends upon a perceptual relation with objects. Such a relationship may obtain between a subject and a manifold of objects, time being a dimension of this manifold, ‘a form of receptivity,’ as Kant puts it.”
Self-consciousness, p. 124.
Receptivity, in this context, connotes an ability to represent an object, apprehend its relevant characteristics, and evaluate it according to identity conditions those characteristics satisfy. For our purposes here, an identity condition can be defined as a finite set of capacities C, which by satisfying an object is designated a common noun N ‘A’. We can formalise this as follows:
A=A ↔ (∃N(A)(C1 ∧ C2 ∧ … ∧ Cn) ∧ (∃x(C1 ∧ C2 ∧ … ∧ Cn)))
If you analyse this formalisation carefully, you'll start to see why the possibility of reference failure becomes relevant in the context of receptive reference. Any capacity C is only satisfied for an object x insofar as at least one of the properties F falling in C’s remit holds for x, as per the formula we came up with in part one:
C(F1, F2,… Fn)→∀x ∃F(F1 ⊻ F2 ⊻ … ⊻ Fn)
Because of this, receptive reference can only hold with absolute certainty if the referer has the necessary epistemic faculties to perceive and recognise every property F capable of being considered in capacity C. Though the above formalisation deals with identity conditions for common nouns, the conclusion here applies a fortiori for proper nouns or names, which are designated by sets of properties F as opposed to capacities C (though this does not entail per se that I agree with the Russellian theory of names as ‘truncated descriptions’). Needless to say, the human being does not have these epistemic faculties. Our ascriptions of both common and proper nouns to objects are not based on exhaustive analyses of whether the object satisfies the set of capacities and/or properties designated by those nouns. Instead, we use a small number of those capacities/properties as shorthand heuristics. Indeed, this is all we can do, as the innate limitations of our sensory faculties preclude us from ever being able to rigorously exhaust the implications of our linguistic practices. Rödl describes the ensuing situation as follows:
“…it does not follow from the nature of this relationship (the relationship between a subject and a manifold of objects perceived by it) that, when I bear it to an object for a stretch of time, I remain connected to the same object throughout. There is space for erroneously thinking that the object has, or has not, been replaced. So I am liable to be confused about the object of receptive reference in these ways: I may be under the illusion of enjoying a continuous receptive relationship with one and the same object, and I may erroneously identify objects I have perceived at different times.”
Self-consciousness, p. 124.
Now, concerning self-consciousness, this manifestly does not hold. There is no room for ‘unnoticed replacements.’ There is no possibility of reference failure, of me failing to get the right object, when I say ‘I’. ‘I’ does not designate any specific finite set of capacities or properties, nor is ‘I’ predicated onto its “object” based on said object satisfying any sort of set of capacities or properties (keep this in mind for later, there’s an important caveat to this). Therefore, ‘I’ does not refer.
“But,” I hear you say, “that surely does not follow. Surely, if I cannot get the object of ‘I’ wrong, then that must mean that I am always getting the object right, not that there is no object whatsoever. There is guaranteed successful reference, not no reference at all.” Ah, but, you see, the concept of guaranteed successful reference is total nonsense— this was, in my humble opinion, decisively demonstrated by G.E.M Anscombe in her (in)famous 1975 essay, The First Person, itself indispensable to Rödl’s programme in Self-consciousness.
To refute this notion, Anscombe asks us to imagine a logician. This logician takes the view that ‘I’, the first person pronoun, is functionally equivalent to a proper name and refers to its intended object in the same manner. To him, if John says ‘I am cold,’ this can be substituted with ‘John is cold’ without altering the validity (or lack thereof) of the claim itself at all— after all, ‘I’ is just John’s name for himself. Of course, our logician recognises that, if ‘I’ refers, as it must in this case, then it must have a form of guaranteed successful reference. But what, in this case, can this guaranteed successful reference actually amount to?
He may say that ‘I’ always successfully refers because, if the object of ‘I’ (the individual using it) did not exist, then the referrer obviously cannot use it (naturally, as they are the individual using it and they do not exist). Thus, if anyone exists and can say ‘I’, then that is enough to guarantee successful reference for their usage of the term. But is this really meaningful reference? If ‘I’ is functionally equivalent to the speaker/referrer’s proper name, then that means that, if you want to know who a given sentence containing ‘I’ holds for if it is true, you need to know the proper name of the individual whose statement the sentence was. If ‘I am cold’ is true insofar as John said it and John is cold, then you need to know that it was John’s statement to know that it was true. This is undoubtedly correct, but where here is there any meaningful reference to an object for ‘I’ qua itself? You cannot argue back from the syntax or grammar of a sentence to infer meaningful reference— nobody thinks that when someone says ‘it is raining’ that ‘it’ is referring to any specific object. Of course, it may be true that it is raining, but that does not at all mean that there is some specific object ‘it’ is referring to that ‘is raining.’ That is reading too much into a simple turn of phrase. Is our logician’s little theory any better?
Our logician, Anscombe then says, might tweak his theory a little bit. Guaranteed reference is not logically tied to ‘I’’s self-referentiality. Guaranteed successful reference amounts to the guaranteed existence of an object when someone uses a name intending to refer to it. This is the type of guaranteed reference that happens when someone uses their own proper name; squaring perfectly with our logicians’s pet theory of ‘I’ being functionally equivalent to proper names all along. Is this it? Not so, Anscombe says. Because there is a very specific sense in which proper names do not have guaranteed reference. Though the existence of the object may be guaranteed, that does not mean it is what I take it to be when I impute a specific name to it. It is entirely possible for one to refer to one’s own name without realising that it refers to oneself. Anscombe illustrates this simply, earlier on in her essay:
“‘When John Smith spoke of James Robinson he was speaking of his brother, but he did not know this.’ That's a possible situation. So similarly is ‘When John Smith spoke of John Horatio Auberon Smith (named in a will perhaps) he was speaking of himself, but he did not know this.’ If so, then 'speaking of' or 'referring to' oneself is compatible with not knowing that the object one speaks of is oneself.”
The First Person, p. 2.
Of course, it is absurd to suppose that such referential difficulties can obtain with regards to the first person pronoun, ‘I’. No one (or at least not an English speaker to whom ‘I’ is not just noise) can say ‘I’ and fail to realise that he is speaking of himself. Thus, ‘I’ cannot be functionally equivalent to proper names. This definition of guaranteed successful reference is also untenable for our poor logician friend.
Having dismissed these potential definitions, Anscombe then moves on to the issue of how successful reference could even be guaranteed in the first place for ‘I’. If we say that the object of ‘I’ is our body, this cannot be the case, as there are numerous circumstances wherein the body is not apparent to us— lucid dreams, hallucinations, states of sensory deprivation, et cetera. None of these circumstances preclude us from using ‘I’— "I’m flying!”, “I feel so peaceful.” and so on are all entirely possible thoughts in these situations. How does guaranteed successful reference still obtain here? If you say that ‘I’ thoughts refer here insofar as these experiential states are induced by one’s bodily neurobiological apparatus, you have landed back to our old logician friend’s first definition of guaranteed successful reference and its attendant issues. Other objectual candidates can be dismissed using similar reasoning. The only suitable candidate, Anscombe avers, would be an immaterial Cartesian Ego. But, as the British government said after deciding to partition off and keep Northern Ireland, “now the troubles start.”
The fact is, attempting to find a discrete object that one can reasonably designate the ‘self’, that could exhaustively integrate the diverse faculties of the mind and body into a single Cartesian Ego, is a fruitless, Sisyphean endeavour. All you may find at the best is a family resemblance between different causal and mental functions. This is what led Hume — and the Buddhists — to conclude that there is no self beyond a mere bundle or aggregation of various properties. There is nothing that a Cartesian Ego could conceivably be. Thus, with the Cartesian Ego dispensed with, we can definitively conclude that ‘I’ does not refer. Does this imply then that ‘I’, or ‘the self’, does not exist, or is somehow a mere illusion? Not at all. Anscombe defines ‘I’ thoughts as consisting of:
…'unmediated agent-or-patient conceptions of actions, happenings and states'. These conceptions are subjectless. That is, they do not involve the connection of what is understood by a predicate with a distinctly conceived subject. (Emphasis mine)
The First Person, p. 16.
The latter part is particularly important. Self-consciousness is entirely spontaneous. It is not conceptually mediated, nor is there an object ascribed a designation on the basis of perceivably satisfying a certain set of properties or capacities. There is no referential object, as simply in virtue of being myself can I designate myself with ‘I’. Thus, the ‘I’-thought, or self-consciousness, is incapable of being predicated onto external objects, as there are no conditions other than being yourself. Thus, then, to simultaneously predicate ‘I’-thoughts to multiple entities violates the law of non-contradiction, as it is equivalent to saying that your individuality qua itself is divisible— which is entirely unintelligible and meaningless.
But that raises another question: we impute self-consciousness onto others, no? Otherwise, how do you avoid your position collapsing into solipsism? Do this, I answer that our theories of mind, the intentional stances we use to consider the internal lives of others, are not direct imputations of self-consciousness onto others. Casually speaking, one's intentional stances vis-a-vis other individuals are extensions of one's own self-consciousness. When one thinks that “If they’re acting like that, they must be angry.” what they are actually thinking is “The most likely cause of my behaviour if I was acting like that would be that I was angry.” Theory of mind is, in essence, a limiting case of self-consciousness. This “other-I”, as I will call it, is a referring term, and its object is the set of behavioural capacities and properties we then impute onto ourselves in a hypothetical sense to produce our theory of mind. And it is here, dear reader, that we come full circle: now, we have the requisite conceptual tools to answer the question that sparked this series: what is death?
Death, we can now see, is a fundamentally linguistic phenomenon. It refers to when the identity conditions required for referral to an “other-I” used to obtain, but do no longer— there are no more any behavioural indicators I can use to construct my theory of mind with regards to that person anymore.
But, if this is what death is, then how does that relate to my death? What happens to my self-consciousness when I die? Well, in short, the same thing that happens when any neurological or physiological change occurs: change, only an infinitely more drastic one than anything experienced beforehand.
To what, though? Beats me. But please stick around for Part 3 if you're interested in a little hypothesis I have.