What is Dennett's view of the function of folk psychology?
Is it a satisfactory alternative to the theory-theory?
In this essay I shall examine Dennett's account of interpretationism and the intentional stance, and how it recasts folk psychology as a predictive strategy rather than as the application of a theory. After critically assessing it and exposing some of its uncertainties I shall examine how it fairs as an alternative to the theory-theory. I shall conclude that, assuming the uncertainties can be clarified, the account makes for a satisfactory alternative.
I shall begin with an overview of the theory-theory.
Folk Psychology and the Theory-Theory
Everyday we make predictions about other peoples' behaviour based on their beliefs and desires. Dave desires beer, Dave believes he has just been paid, therefore I predict Dave will go to the pub. This is everyday, commonsense, folk psychology. The theory-theory is the philosophical theory that, when we make these everyday psychological predictions, we are utilising an empirical theory, a theory referred to as "FP". It may not be explicitly formulated or consciously used, but it is tacitly understood (Frankish 2005).
Intentional states - beliefs, desires etc. - are to FP what atoms and molecules are to chemistry. They are the fundamental theoretical building blocks upon which the theory rests. Their definition, and indeed the very truth of their existence, is bound up in the theory of which they are a part. FP is an empirical theory and therefore could, like any empirical theory, turn out to be radically false. Should this happen, it is claimed, the intentional states that it defines cannot exist. They will go the way of phlogiston and calorific.
Eliminative materialism is the thesis that this is exactly what will happen. FP will be usurped by a modern, materialist theory presumably derived from neuroscience, and intentional states will be shown to be antiquated fictions. By contrast, Realism is the thesis that intentional states exist, and any theory of neuroscience will demonstrate their existence, for example by mapping individual beliefs onto specific neural states.
Neuroscience has yet to advance sufficiently for us to know which of these theses is likely to win out, but if the theory-theory is correct, and eliminative materialism is correct, then our entire way of relating to each other as people will be overturned.
Dennett presents a slightly different view of folk psychology. If it can be shown to be a satisfactory alternative to the theory-theory we can head off this potential upheaval before it happens. We'll examine his view now.
Interpretationism and the Intentional Stance
Rather than talking about utilising the theory of FP, Dennett instead talks about adopting the intentional stance. When faced with a system (a plant, a device, a person) and attempting to predict its behaviour in a given situation there are a number of strategies - or stances - one can adopt. Most fundamental is the physical stance, whereby one studies the physical composition of the system - down to the atomic level if necessary - and predicts behaviour based on the laws of physics. While this strategy is accurate and reliable, it can be prohibitively complex in large systems.
Another strategy is the design stance. Here one studies the functions of the system to ascertain what it should do in the circumstances if it is correctly fulfilling its purpose.
Lastly we could adopt the intentional stance. Here one posits that the system is a rational agent possessing intentional states. One ascertains all the relevant intentional states possessed, applies practical reasoning, and predicts what the system ought to do if it were rational. This stance can be quite accurate and for extremely complex systems - such as humans - can be the only stance that is practical (Dennett 1987).
Thus while the theory-theory says that folk psychology is the application of theory, Dennett is saying it is the application of a strategy. A theory stands or falls on its truth. A strategy stands or falls on its success, and we know the intentional stance is a successful strategy as we use it everyday. Regardless of what neuroscience may or may not discover in the future, the intentional stance will remain viable.
Let's now examine some criticisms of Dennett.
Criticising Dennett - Is it an alternative?
One of the strengths of Dennett's idea is that it is neutral to the reality of intentional states. As part of an interpretative strategy they need not be physically real, but merely instrumentally real. Rather than being compared to atoms and molecules, they can better be compared to centres of gravity or the square root of minus one.
However, with a relatively simple move we can achieve the same effect with FP. We simply recast intentional states as "functional states". Functionalism has a similar substance neutrality, in that a belief is defined by its function rather than by its composition, and a function is also only instrumentally real.
However, there is a crucial difference. Functionalism may recast the nature of its intentional states, but it still remains an empirical theory and thus stands or falls on truth rather than success. Compare the following statements:-
- I theorise that my teaspoon has the function of cutting toast.
- I shall apply the strategy of cutting toast with my teaspoon
Both statements posit an identical "cutting toast" function, but statement 1) can be empirically false whereas statement 2) can merely be unsuccessful. It's conceivable that I've enjoyed a long and successful history of cutting toast with my teaspoon even while statement 1) is false.
This shows that what makes Dennett's position a promising alternative to the theory-theory is not replacing "real" with "instrumentally real", but replacing "theory" with "strategy". Let's consider now how satisfactory an alternative it is.
Criticising Dennett - Is it satisfactory?
We've seen that under Dennett's account intentional states are merely interpretively useful. A problem arises when we realise that they may be usefully applied to interpret systems that are clearly inanimate.
Dennett says that to possess intentional states, to be a "true believer", is to be "a system whose behaviour is reliably and voluminously predictable via the intentional strategy" (Dennett 1987). However, this opens the door to the criticism that anything can have beliefs. Consider my laptop. Despite being a more than competent user, I often find it useful to apply the intentional stance in order to explain its otherwise unfathomable behaviour. The implication is that if I can refine my application of the intentional stance to the point where it "reliably and voluminously" predicts the behaviour of my laptop, then my laptop is a true believer.
Dennett addresses this objection by claiming that applying the intentional stance to true believers will reveal patterns of behaviour that are invisible to any other strategy. Applying the intentional stance to my laptop may well be effective but will reveal nothing that couldn't in principle be revealed by applying the physical stance, and therefore it is not a true believer. By contrast, applying the physical stance to humans would miss the behavioural patterns we exhibit (Dennett, 1987, pp7-11).
Dennett's account hangs on this idea, but precisely what he means is not clear (Barber 2005, p164). Humans are very adept at seeing patterns where none exist (shapes in the clouds, Rorschach ink-blots, etc.). If I do see behavioural patterns in my laptop how do I know if they are real or imagined? If the intentional stance is to be a satisfactory alternative to the theory-theory, we need to clarify this.
Dennett addresses this issue at length in his 1991 paper. He suggests that a pattern in a system is real if it is more efficient to describe the system with reference to the pattern than describing the system in its entirety (Dennett, 1991, p8). So, if I could describe a cloud fully by saying "it's a cloud that looks like a sheep" this would be more efficient than a full description of all the various lumps, curls and wisps, and therefore the sheep pattern is real.
However, it is quite possible for the same system to be described in reference to two entirely different patterns. The cloud might look like a sheep to me, and a dog to someone else. This remains true if we refer back to people. Imagine two observers are observing me as I scowl at a half-empty glass of gin & tonic and then take it into the kitchen. Applying the intentional stance, Observer A concludes that I desire a top-up and have gone for a refill. Observer B concludes that I believe I have drunk too much and have gone to tip the remainder away. Both see a "real" behavioural pattern, but they cannot both be right.
This example highlights another problem for Dennett's account. If beliefs are no more than the posits of a predictive strategy, then they require someone to do the predicting. Plus, if different predictors posit contradictory beliefs, which are correct?
This objection could be deflected somewhat if we regard the person themself as a privileged predictor. I can thus posit my own beliefs, and in the event of a discrepancy with another predictor, my predictions take precedence. Curiously though, in a 1990 paper, Dennet explicitly denies this. He advocates a variation of the "intentional fallacy", which has its roots in Literary theory and claims that an utterance can have no unequivocally correct interpretation, even if it's the utterer doing the interpreting.
For Dennett's account to be satisfactory, the vagueness of patterns needs to be addressed, and an explanation needs to be provided to deal with contradictory posited beliefs. However, I do not regard either of these problems to be fatal flaws in the account.
Criticising Dennett - Is it a satisfactory alternative?
To close, I shall examine some attacks levelled at the theory-theory and assess whether Dennett's interpretationism can provide a defence.
Paul Churchland claims that recasting outdated fictions as instrumental constructions is a strategy that alchemists could have adopted to save their own patently false ideas. Applying this tactic to intentional states is, he claims, equally disreputable (Churchland 1981).
Doubtless the alchemist could have tried to save his theory in this way, but he would only have succeeded if such a recasting proved useful. Alchemical theories were usurped because the new theories were demonstrably more effective. Neuroscience may well produce a theory demonstrably more effective than FP and thus usurp it, but the intentional stance will remain useful. Neuroscience is not a science for everyday use. What's likely is that a neurological stance will emerge and take its place between the physical stance and the design stance, leaving the intentional stance unaffected.
Stephen Stich argues that one must either be a Realist or an Eliminativist towards intentional states, but that instrumental or functional realism is untenable (Stich 1988). He considers the Connectionist neurological model - although he could equally well use a hypothetical Theory X - and argues that there are only three possibilities:-
- FP is true. Thus, intentional states are wholly real.
- Connectionism is true, but Connectionism can accommodate intentional states. Thus, they are still wholly real.
- As scenario 2, but Connectionism cannot accommodate intentional states. Thus, they are wholly unreal.
Therefore, intentional states are either entirely real or entirely unreal, but not instrumentally real.
There are a couple of replies to Stich. Firstly, he fails to consider a fourth case:
- Connectionism is true, and Connectionism can partly accommodate intentional states.
Perhaps intentional states can be mapped to approximations of neural states in the way that Newtonian physics is an approximation of Einsteinian physics. Perhaps we discover a kind of mental/neural duality akin to the wave/particle duality of quantum mechanics.
More relevantly for us, the truth or otherwise of Connectionism is unlikely to effect the intentional strategy. The unconsidered fourth alternative would now be:
- Connectionism is true, but is too complex for everyday use. The intentional stance remains the most usable predictive strategy.
I have only considered two possible lines of attack, but I assert that others will succumb to similar defences. However, we have exposed the primary weakness of Dennett's account.
The intentional stance is a predictive strategy that stands or falls on its success. For it to fall, all that is required is for something more successful to come along, be it predictive strategy, Connectionist theory or alchemical black art. As things stand though, there are no signs of any such threats.
Conclusions
Dennett presents a viable view of folk psychology as the application of the intentional stance, as compared to the application of the theory of FP. Areas of his account require clarification but, assuming suitable clarifications can be made, the account is a satisfactory alternative to the theory-theory.
Bibliography
- Aitchison, Jean 1992. "Linguistics", (Sevenoaks: Hodder & Stoughton), 4th edn.
- Barber, Alex 2005. "Language and Thought", (Milton Keynes: Open University), ch7, pp155-167
- Churchland, P.M. 1981. "Eliminative materialism and the propositional attitudes". Journal of Philosophy, vol. 78, no.2, pp67-90
- Churchland, P. M. 1988. "The ontological status of intentional states: nailing Folk Psychology to its perch", Behavioural and Brain Sciences, vol.11, no. 3, pp507-8
- Dennett, Daniel 1987. "True believers: The intentional strategy and why it works" in Dennett, D.C. 1987. The Intentional Stance (Cambridge MA: MIT Press). pp13-35.
- Dennett, Daniel 1990. "The Interpretation of Texts, People and Other Artifacts", Philosophy and Phenomenological Research, Vol 50, Supplement (Autumn, 1990), pp177-194
- Dennett, Daniel 1991. "Real Patterns", The Journal of Philosophy, Vol 88, No. 1, (Jan., 1991), pp27-51
- Frankish, Keith 2005. "Folk Psychology: The Last Myth?", The postgraduate foundation module in philosophy, (Milton Keynes: Open University), pp152-233
- Stanford Encyclopaedia of Philosophy, <http://www.seop.leeds.ac.uk/entries/folkpsych-theory/> [accessed 28th July 2006]
- Stich, Stephen p. 1985. "The future of folk psychology", From Folk Psychology to Cognitive Science, (Cambride MA: MIT Press) pp242-246, and p253
- Stich, Stephen P. 1988. "Connectionism, Realism and realism", Behavioural and Brain Sciences. vol. 11, no. 3, pp531-2
