Explain and evaluate Block's 'homunculi-headed robots' (Chinese nation) objection to functionalism. Do you think that the objection in question is a good one? Consider possible functionalist replies to it.

Ned Block's homunculi-headed robots objection to functionalism is an attack on functionalism's liberal tendencies. By 'liberal', Block is referring to functionalism's classification of certain systems that lack mental abilities as, in fact, possessing mental abilities (Block 2002, p. 96). In what follows I will provide a brief outline of what a homunculus-head is and I will evaluate whether the homunculus-head is an effective approach to counter functionalism.

Functionalism is the theory that a given mental state, a pain for example, is identical to a particular functional state, being causally dependant upon the functional state's inputs (senses), outputs (behaviours), and other internal (mental) states (Block 2002, p. 94; Levin 2004). Further, Machine Functionalism¹ claims that all mental states are analogous to given machine-table states (Block 2002, p. 97). The example of the Turing Machine is often used to illustrate this point and Block outlines one such system, the homunculus-head. To understand Block's homunculus-head imagine a robot externally similar to a human but internally drastically different. Instead of nerves picture a series of wires that when the sense organs are stimulated turn on certain lights inside the robot. To control the muscles, and hence the robot's behaviour, there are a series of buttons inside the robot too. Also, inside the robot, is a machine table that describes the 'person'² that is the robot and a man that is trained to interpret the light patterns triggered by the sense organs by referring to the machine table and pushing the corresponding buttons to cause a particular behaviour. An example of this homunculus-head in operation would be as follows. The man inside the robot looks to the lights to see what current inputs are occurring; it turns out that the man has been trained to recognise the lights that are illuminated as pattern I₁₂₃₄₅. The man turns to the machine table and finds that for I₁₂₃₄₅ the corresponding output is O₅₄₃₂₁ which the man knows to achieve by pressing a certain combination of the buttons and hence the output dictated by the machine table is achieved (Block 1980, p. 260). It is evident how such a system is functionally correlative to a 'person'; the external behaviours would seem identical.

In Troubles With Functionalism (pp. 96-97) Block provides two examples of such a homunculus-head with the first being an abstract body not unlike our own and the second being an abstract application of the population of the nation of China. Block uses the Chinese nation as an example of a homunculi-head because the population of China is roughly the same as the amount of neurons in the brain, one billion (Block 2002, p. 96). We can say that each of the inhabitants of China is the equivalent of the man in the example given above but are connected to the robot's body via two-way radios instead of lights and buttons. Also, the machine table is displayed on a series of strategically placed satellites visibile from any location within China (Block 2002, p. 96). The reason Block gives the example of the Chinese nation as a possible homunculus-head is that while the first example above may be metaphysically possible, the abstract application of the population of China as a homunculi-head is both nomologically and physically possible (Block 2002, p.96).

At this point it may be asked what Block is trying to argue? This illustration of functionalism seems to outline 'persons' quite well. Block's point, however, is in what is lacking; qualia. The reason Block claims the homunculus-head is a valid counter-consideration to functionalism is because there is a certain amount of ambiguity as to whether the homunculus-head not only has mental states but whether the homunculus-head has 'qualitative mental states' or qualia (Block 1980, p. 261). To illustrate this let us use the example of pain and imagine two 'people', one is yourself and the other is a homunculi-head. You are currently in pain because you have twisted your ankle and are therefore crying. The homunculi-head is currently in the same functional state as you and as the population of China who are currently controlling the homunculi-head are interpreting the machine table for this particular functional state correctly the homunculi-head is also crying. To an external observer both you and the homunculi-head appear identical, you both appear in pain from similar injuries. However there is no reason to conclude that the homunculi-head is actually in pain; that is, to be experiencing the quality of pain or what it is like to be in pain (Block 1980, p. 261). Block calls this the 'absent qualia argument' (Block 2002,p. 97).

Block's argument can therefore be summarised as follows. Functionalism, or more specifically Machine Functionalism, claims that all mental states are analogous to machine-table states (Block 2002, p. 97). Block (2002, p. 97) gives the following example; 'a particular qualitative state, Q, is identical to a machine-table state, Sq. But if there is nothing it is like to be the homunculi-headed system, it cannot be in Q even when it is in Sq'. If Block's position holds then Q can never equal Sq and functionalism, or at least Machine Functionalism, cannot, therefore, hold.

In Troubles With Functionalism (2002, pp. 96-97) Block provides a number of counter-considerations to his Chinese nation example of a homunculi-headed system. These counter-considerations all relate to the nomological relevance and possibility of such a system and, being quite weak arguments, are refuted by Block quite effectively. Perhaps the strongest counter-consideration, however, against Block's homunculus-head argument against functionalism is Sydney Shoemaker's Absent Qualia are Impossible--A Reply to Block. Shoemaker argues that Block's definition of functionalism is over simplified and that when functionalism is defined adequately Block's homunculus-head argument becomes irrelevant (Shoemaker 1981, pp. 587, 590-591).

Shoemaker agrees, in principle, with Block's simplified definition of functionalism that functionalism equates a qualitative state (Block uses pain as his example) with a given functional state that is definable through its causal relationship to inputs, outputs, and other mental states (Shoemaker 1981, p. 582). Shoemaker, however, does have an issue with what is meant by other mental states; it is possible, for example, to conceive of a mental state that may be difficult to define functionally and Shoemaker uses the example of beliefs to illustrate this. Block's definition of functionalism, however, appears to have no prima facie antagonism with the definition of other mental states; Block simply claims that a certain qualitative state, be it a pain or something else, only needs to have a causal relationship to other mental states, not that the other mental states be functionally definable (Shoemaker 1981, p. 582). Shoemaker claims, however, that Block removes all mental terminology from the functional definition of pain which would entail that the other mental states that pain is causally related to would likewise need to be functionally definable; Block also makes reference to this in his footnote to the above definition of functionalism (Shoemaker 1981, p. 582).

Shoemaker proposes a variant definition of functionalism that has two variables consistent with Block's position. These two variants of the same definition focus on different interpretations of the possibility of a functional definition of the other mental states mentioned in Block's definition of functionalism presented above. The first variant of Shoemaker's definition is known as 'Absent Qualia Thesis One' (AQT-1) and claims that qualitative states (mental states) are not functionally definable at all (Shoemaker 1981, pp. 582-584). If AQT-1 held it would be possible for a 'person' to have a genuine mental state (although non-qualitative), such as a belief that they were in pain, even if the pain were an imitation pain and thus lacked qualitative character. 'Absent Qualia Thesis 2' (AQT-2) differs from AQT-1 in that where AQT-1 doesn't require any functional definition of mental states at all, AQT-2 requires what Shoemaker calls a 'weak functional definition' (Shoemaker 1981, pp. 582-583). This weak functional definition is the same as Block's primary definition of functionalism where it is irrelevant whether mental states exist in relation to functionally definable mental states or functionally undefinable mental states (Shoemaker 1981, p. 582-583). The converse of the weak functional definition is the strong functional definition which is analogous to what Block outlines in his footnote relating to his outline of functionalism. In his footnote Block uses the Ramsey sentence to remove any mental terminology from definitions of qualitative states so for any qualitative state to hold, according to the strong functional definition, it must be definable in such a way (Shoemaker 1981, p. 583).

It may seem unclear as to how this outline of definitions by Shoemaker is relevant to Block's homunculus-head objection to functionalism, however, on closer inspection the importance becomes evident. Shoemaker argues that functionalism should be understood in relation to the strong functional definition outlined above however Block's homunculus-head objection focuses solely, according to Shoemaker, on refuting AQT-1 (Shoemaker 1981, pp. 583, 587). While AQT-1 being false certainly weakens functionalism's claims, the fact that Block doesn't attempt to counter AQT-2 is insightful. As Shoemaker (1981, p. 587) points out, '[f]or anyone who holds AQT-2, but denies AQT-1, it will be difficult to maintain that qualia are the achilles heel of functionalism'.

Block's homunculus-head objection to functionalism can be seen as effective so long as it is applied to a limited reading of functionalism. If Block's homunculus-head objection is presented as a refutation solely of machine functionalism, when machine functionalism is in AQT-1, then this can be seen as effective and important in marking the limits of functionalism. However, if Block's homunculus-head objection is applied more broadly, particularly in relation to Shoemaker's different definitions of functionalism, then the weaknesses of the argument become apparent. For Block's homunculus-head objection to functionalism to have more universal relevance within a broader definition of functionalism a wider definition of mental states and their relationship to other mental states is necessary.

1601 Words.

1. Machine Functionalism is the 'type' of Functionalism that Ned Block is targeting in his paper Troubles With Functionalism.

2. In what follows, 'person', when presented in apostrophes, will denote functionalism's definition of person.

References

Block, N 2002, 'Troubles With Functionalism', in David Chalmers, Philosophy of Mind: Classical and Contemporary Readings, Oxford University Press, New York, pp. 94-98.

Block, N 1980, Are Absent Qualia Impossible?, The Philosophical Review, Vol. 89, No. 2, pp. 257-274.

Levin, J 2004, 'Functionalism', in Stanford Encyclopedia of Philosophy, <http://setis.library.usyd.edu.au/stanford/entries/functionalism/#1>.

Shoemaker, S 1981, 'Absent Qualia are Impossible—A Reply to Block', The Philosophical Review, Vol, 90, No. 4, pp. 581-599.