Herein I define the essential structure of Putnam’s Twin Earth Argument; describe the conclusions it entails on the position of meaning and reference in natural language; show how Stich adopts this position to argue against beliefs as explanatory devices within the framework of psychology; and explain the ways in which Stich’s critique is both successful and unsuccessful.

Putnam begins by describing the historical situation :

1. Knowing the meaning of a term is being in a certain psychological state.
2. That the meaning of a term determines its extension, e.g to what things the term applies.

Even when meaning is thought of as a public thing, as for example Frege would claim, to know the meaning of a concept in those cases still would rely on some sort of psychological state. However, Putnam believes this historical basis is wrong, essentially because he is against the idea that being in a certain psychological state can define the set of things that the referent of such a state is believed to apply to. He gives a thought experiment, known as the Twin Earth Argument, to counter the historical basis :

• Imagine a twin earth TEarth, an exact replica of Earth in every detail, including people.
• Except, water on TEarth is not H20, but a more complex formula refered to as XYZ.
• Further, take the time to be in the 1700s for both Earth and TEarth – therefore people do not yet know about H20 and XYZ.
• Hence, the meaning of water to Earth person Omar and a TEarth person TOmar would seem to be the same – there would be no known way to distinguish water from Twater.
• But in 50 odd years, Omar and TOmar will discover that in fact water and Twater are not the same.

Further examples are given in varying contexts including ones solely on Earh. Putnam claims then that just knowing the meaning of water cannot determine to what water refers – for initially Omar and TOmar would refer to two sets of things as water, later discovering they were different, and that they had been referring to different sets of things all along. Hence, according to Putnam, it is impossible to define the meaning of a term based solely on what psychological state may be instantiated. This view has come to be known as semantic externalism.

Clearly, this argument relies on the idea that Omar and TOmar do in fact have the same psychological states when they think about water, whereas it has been argued this may not be the case. Also, even if it does perhaps show that Omar and TOmar are referring to different things, it also assumes a link between meaning and reference – that if a term has a different reference, it cannot then have the same meaning. Otherwise, the conclusion would be that references are not in the head.

Putnam prefers instead to rely on the concepts of causal history and the division of linguistic labour to explain meaning. Experts may know exactly what gold is but it is unnecessary and inefficient for everyone to know, or to be able to identify it, yet on the whole the word has meaning. He concludes that without this reliance on causal history and division of linguistic labour, meaning cannot be discerned.

One could counterargue that water is not a rigid designator (as defined by Saul Kripke) for H2O, but only a term that applies in some cases where it is useful to utilise that application. if it is not a rigid designator, putnam could not necessarily show that a person on earth thinking about water was in fact referring to H2O. Also, it is questionable whether semantic reduction must take place – it may be possible to use the term water without intending to refer to H2O atall, as Searle would claim.

Regardless of the objections made to the Twin Earth Argument, Stich adopts Putnam’s position to argue against beliefs as explanatory devices within the framework of psychology. It is often claimed, that human actions can be explained in terms of beliefs and desires (the belief-desire thesis) and furthermore that this thesis is required for all modern concepts for cause and action, to which Stich would agree. However he believes there is inconsistency between the belief-desire thesis and what he describes as the principle of psychological autonomy. He proposes to show that many belief-desire explanations of action cannot allow for autonomy and that sacrifices must be made to resolve this conflict.

The principle of psychological autonomy :

• Imagine we can build perfect, exact copies of people
• The original and the copy will therefore also be psychologically identical – any psychological instantiation occurring in one will occur in the other
• Further, stich is not worried at what times they both exist, or if they are perfect copies on purpose or by accident – they will still instantiate the same psychological properties
• Also, the shared psychological properties are of the form “I believe X”. Therefore the possible fact that X is not true is irrelevant for the principle. The psychological property believing X is what is at stake, rather than the property of knowing X

The belief-desire thesis :

• Human actions can be explained, if not completely then partially, by beliefs and desires
• For current purposes, these should be seen as literal causes for actions
• Also, the statement “X did Y for believing Z” should be accepted as a singular causal statement more generally explainable (but currently unnecessarily) by laws of relations between X, Y and Z
• If a belief A is true, and belief B is false, the beliefs are of different types

With these two principles clearly defined, Stich provides 4 example cases of belief types in order to reveal the inconsistency he alludes to. Here are the formats for the examples :

1. Self-referential beliefs

• X did Y for believing Z
• Z must be considered an explanatory belief for causing Z within the belief-desire thesis
• However, X2 (an exact copy of X as per the principle of psychological autonomy) comes into existence and does Y2 for believing Z2
• But if (as per Stich’s example) X2 has just been created and answers a question about a past experience, Z2 is necessarily false, as X2 has no past experiences.
• If Z2 is false, it cannot be the same belief type as Z
• Thus Z cannot be an explanatory belief

2. Beliefs about spatial and temporal location

• Essentially the same as 1, except instead of being copied, X is frozen and thawed in the future.
• At the future time, the belief Z has become false
• The belief Z is therefore not the same belief as it was in the past, yet may cause the same Y to take place, e.g. hunting for strawberries in a place where strawberries once grew
• Thus, again, Z cannot be an explanatory belief

3. Beliefs about other people

• Invokes Putnam’s Twin Earth Argument
• But X believes Z about a person (and it is true) rather than about water
• And TX holds a belief TZ about a person
• But belief TZ is false – on TEarth, it is not true for that Tperson
• So Z and TZ cannot be the same belief type
• Therefore, again, Z cannot be an explanatory belief

4. Natural kind predicates

• Again invokes the Twin Earth Argument
• This time a person believes something that is untrue on Earth – that lizards dissolve in water
• But that is true in TEarth – as water there is actually XYZ and can dissolve said lizard
• As previous examples then, Z cannot be an explanatory belief

Thus Stich concludes that whilst the belief-desire thesis is not false, it is the case that, if the principle of psychological autonomy is also true, then there are a large set of beliefs that cannot be causes in an explanatory theory of psychology. Stich believes that this, assuming one accepts the principle of psychological autonomy, shows the belief-desire thesis to be very much less applicable in causal explanation of action.

One could claim that Stich’s argument relies too heavily on the very clear definition he gives of the types of beliefs he wishes to allow into the belief-desire thesis. However Stich gives a reply to such a claim – which is essentially that, even if the types he describes – known as non-autonomous, hybrid or de re beliefs – are supplemented by autonomous beliefs (of which the non-autonomous are hybrids), the belief-desire theorist is still not much better off. For there is as yet no clear argument for the distinction between non-autonomous and autonomous beliefs and any distinction sufficient to allow autonomous beliefs to be free from his inconsistency may cause them to cease being beliefs in any sense of the word – they would no longer refer to or represent anything, and could hold no truth value.

It is clear that Stich takes Putnam’s position that meaning is not resultant of psychological state, and reformulates the thought experiment along with the principle of psychological autonomy to illuminate the inconsistencies of the belief-desire thesis. Whilst thought experiments by their nature can require complex aspects of an argument to be over-simplified, the underlying structure of Stich’s four examples does raise the question, for any theories reliant on the concept of causal beliefs and desires, of how to either implement them with, or formulate them as independent of, psychological autonomy.

Stich provides a good argument that highlights inconsistencies in the framework of explanatory psychology, but acknowledges his inability to disprove the belief-desire thesis. However, he does succeed in leaving the belief-desire theorist with a very heavy burden of proof.

REFERENCES

• Fodor, J A. (1975) The Language of Thought, Harvard University Press
• Kripke, S. (1972) Naming and Necessity, In D. Davidson and G. Harman (eds) Semantics and Natural Language, Reidel
• Putnam, H. (Nov. 8. 1973) Meaning and Reference, The Journal of Philosophy, Vol. 70, No. 19, Seventieth Annual Meeting of the American Philosophical Association Eastern Division, pp. 699 – 711
• http://links.jstor.org/sici?sici=0022-362X%2819731108%2970%3A19%3C699%3AMAR%3E2.0.CO%3B2-1
• Stich, S.P (1993) Autonomous Psychology and the Belief-Desire Thesis, In Goldman A.I (ed) Readings in Philosophy and Cognitive Science, MIT Press