9. Bibliography
2.6. The Spectre Argument
2.6.1. What Are Spectres?
A close cousin of the zombie argument is the spectre argument. This differs from the zombie argument only in that it deals with a situation in which phenomenal properties are not absent but in some way different from those in the actual world. The classical (and still most common) example of this is the inverted spectrum case, which dates back to Locke, and has been recently presented by Shoemaker (1982),14 although it is in Chalmers (1996) that it is aimed specifically at physicalism.
14 Shoemaker’s target was functionalism, and in particular representationalism.
In an inverted spectrum scenario, when I look at a ripe tomato I experience red qualia and say “that’s red”, but when my qualia-inverted physical duplicate looks at a ripe tomato he experiences green qualia and says “that’s red”. We have learned to call the ripe tomato experience “red”, even though what we actually experience is
different. We are physically the same but phenomenally different, and this is the minimal requirement for a scenario to provide a counter-example to physicalism.
Furthermore, there are simpler scenarios (other than inversion) which satisfy this requirement. Any change in phenomenal properties that is not accompanied by a change in physical properties is enough to serve as a counter-example to physicalism. Instead of inverting qualia space, we need merely stretch or shrink it. We could, for example, tinker with the intensity of pains and pleasures. They would all still feel bad and good, respectively, and so play their rightful roles in the overall cognitive
economy, but the pains would feel less bad and the pleasures feel less good. Since these balance out, it is plausible that the overall behaviour would not be effected. Likewise we could adjust the saturation of my duplicate’s colour qualia space, so that his reds would still be reds, but they would be less intense than my reds. However, the intensity relations amongst his qualia would be indiscernible from mine. This is, I think, the most modest sort of phenomenal difference, and so this “low-intensity
spectre” scenario is the easiest to defend. I will discuss this, and other kinds of spectres, in chapter seven.
As I did with zombies, let me offer a precise definition of spectres. Since being a spectre is defined in relation to some other being, we can talk about a spectre-
duplicate.
x is a spectre duplicate ofy if and only if:
(1) x is a minimal physical duplicate ofy, and
(2) y is experiencing phenomenal consciousness of some specific type, and
(3) x is experiencing phenomenal consciousness of some different type. Notice that, unlike the definition of zombies, the relation between the
duplicates is symmetrical. It seems odd to say that they are both spectres, so let us allow thaty is experiencing “normal” consciousness, and x is the spectre. I will use the word “spectre” to refer to all creatures which have physical sameness and phenomenal difference from stipulated “normal” creatures. Where the difference is complete (e.g. visual spectral) inversion, I will call the beast an “inverted spectre”. In
general I will refer to spectres and the spectre argument, meaning any kind of phenomenal difference, but I will generally have in mind the kind of “low-intensity” spectre mentioned above. For what it is worth, zombies might also be considered a special type of spectre, a “zero-intensity” spectre, but since zombies illustrate the general problem and all other spectres illustrate the specific problem, it is better to stipulate that zombies aren’t spectres, and proceed.
2.6.2. The Spectre Argument Formulated
The spectre argument could be expressed using the same seven premises used for the zombie argument, making only minor alterations to (1), (2), and (4), as follows.
(1) I am experiencing phenomenal consciousness of a specific type, (premise) (2) It is conceivable that there is a world, w, such that w is a minimal physical duplicate of the actual world and I am experiencing phenomenal consciousness of some different type at w. (premise)
(3) Conceivability is a guide to possibility, (premise).
(4) There is a possible world, w, such that w is a minimal physical duplicate of the actual world and I am experiencing phenomenal consciousness of some different type at w. (from 2, 3).
(5) There is a possible world, w, such that w is a minimal physical duplicate of the actual world, and not a duplicate simpliciter of the actual world, (from 1, 4).
(6) If there is a possible world, w, such that w is a minimal physical duplicate of the actual world and not a duplicate simpliciter of the actual world, then
physicalism is false at the actual world, (from the definition of physicalism). (7) Physicalism is false at the actual world, (from 5, 6).
This argument, like the zombie argument, seeks to show that there is a possible world which is a minimal physical duplicate of the actual world but not a duplicate simpliciter of the actual world. Instead of there being something missing, there is something different.
It should be noted that the conceivability of spectres is relatively safe to the sorts of objections made against the conceivability of zombies. These objection use the fact that phenomenal properties are necessary for normal cognitive functioning, and so duplicates must have them.
However, spectres face no such problems. They definitely do have
only differ from the actual world in that they have different properties, rather than none at all.
Another objection to the conceivability of inverted spectres is based on the claim that we cannot systematically invert qualia and maintain the same behaviour. This is clearest in the case of pleasure and pain. Can we really imagine a case of a duplicate who puts their hand over a flame, says “ouch”, removes their hand, and yet feels pleasure (inverted pain) as they do so?
However, while the pleasure/pain example seems to resist inversion, other qualia seem more amenable to inversion. The trump-card here is of course than inversion is not necessary; the more modest requirement of shrinking the qualia space will do. The conceivability of all kinds of spectres will be discussed in chapter seven. I will argue that spectres are, in some sense, conceivable.
I will respond to the spectre argument by disputing the inference from conceivability to possibility as expressed in (3). I will argue that conceivability evidence is unreliable in cases like spectres, because of some special features of phenomenal concepts. So the sort of conceivability involved in conceiving of spectres does not warrant the inference to possibility. I will examine these issues in chapter seven.