Descartes once described thought as a “universal instrument which can be used in all kinds of situations.”
The fact that thought represents objects in a “disengaged” and “stimulus-independent” manner allows us to think about objects in their absence. We can think about events that have not yet occurred and we can think about events that will never occur; indeed, we can think about events that could never occur. This capacity allows us to anticipate the consequences of certain events in advance of their happening and to prepare for them. If the anticipated consequences of an event are positive then one might attempt to bring it about; and if those consequences are negative then one can take steps to prevent it from occurring. Thus, a creature with the capacity for thought can control its environment in a way that a creature that is reliant only on perception cannot.
A second feature of thought is its scope. Whereas perception provides us with access to only a small range of things, the reach of thought is (practically) unlimited. We cannot see objects that are very small, we cannot hear sounds that are very high in pitch, and we cannot detect odors that are very faint. But there is no such limit on the range of objects that we can think about.
Thought enables us to grasp features of the world that are not perceptually accessible to us. A creature that possesses only perceptual capacities can respond to the physical features of its environment, but it cannot respond to its economic, political, or psychological features. It cannot take steps to nullify the effects of inflation, participate in an election, or attempt to convince someone that their views are mistaken.
Getting a joke typically requires appreciating connections between topics that are normally unrelated.
The idea here is that although one can visually identify an object as a cat without possessing the concept of a cat, one cannot think about a cat without possessing the concept of a cat.
In order to understand thinking we need to understand the ways in which thoughts are related to each other. Although thoughts can occur in isolation — while waiting at a traffic light one might suddenly be struck by the thought that friendship is a fundamental goal — it is perhaps more common for thoughts to occur as components of trains of thoughts, sequences of thoughts that are related to each other in some way.
There are 2 ways in which the components of a train of thought can be related to each other. Some trains of thoughts involve only associative relations — thoughts introduce each other with a certain degree of methods and regularity. Associative thinking is familiar from daydreams and other forms of reverie. One begins by wondering whether films that are based on books are in general as good as the books on which they are based, which leads one to wonder what films have recently been released, which in turn causes one to wonder what day of the week it is, which reminds one of a deadline, and so on.
Trains of thoughts can be linked by inferential relations. Consider the thoughts “Socrates is a human,” “All humans are mortal” and “Socrates is mortal.” The first 2 thoughts provide one with reasons to endorse the 3rd thought.
Where should we look for principles that might govern the evaluation and regulation of thought? Many theorists have suggested that we should look to the formal systems of logic and probability theory here. But although logic and probability theory provide us with some pointers as to how we ought to think, their usefulness in this regard is surprisingly limited. For one thing, at best logic and probability theory tell one what one should not think rather than what one should think.
A second limitation of logic and probability theory is that neither discipline takes into account the constraints under which we must think. Thinking is always situated in a particular context, and what counts as good thinking depends on the constraints that govern the relevant context. In some contexts one has all the time in the world to think through a problem, and the need to get the right answer is more important than getting a speedy answer. In other contexts time is of the essence, and a rough approximation to the truth that is arrived at quickly might be better than a perfectly correct verdict which arrives too late to be of any use.
The brain has muscles of thinking as the legs have muscles for walking.
We can think of CTT as comprised of 2 claims: a claim about the nature of thoughts and a claim about the nature of thinking. In a nutshell, CTT holds that thoughts are sentences in a “language of thought,” and it claims that thinking involves formally governed transitions between sentences in a language of thought.
Instead, one could think that the structure of thought has more in common with that of maps or diagrams and is not particularly language-like. In other words, the advocate of CTT might suggest that the reference to the “language” of thought should be taken with a large grain of salt. What CTT does require is that thought has a compositional structure, such that the parts of a thought make independent contributions to its meaning.
That, in essence, is the computational theory of thought. In a nutshell, the idea is that thoughts are symbol structures, and thinking involves the manipulation of these structures on the basis of their formal properties. Of course, it is not clear that thinking feels like symbol manipulation, but CTT does not make any claims about what it feels like to think. Rather, CTT is an account of how purely physical creatures could be thinkers.
Searle asks you to suppose that you are located in a room into which Chinese messages are sent. Although you do not understand the messages, you have access to a giant “look-up” table that maps each of the messages that you receive to a suitable response.
Searle claims that although someone in the Chinese room would be able to manipulate symbols in the appropriate manner, they would not be thinking. What this show is that mere symbol manipulation does not suffice for thought — genuine thought requires something more. But since CTT equates thought with symbol manipulation it must be false.
Symbols are manipulated purely on the basis of their formal properties — the whole point of the account is that they do not have to be understood in order to be manipulated. The appropriate analogue of the thinker is not the person in the Chinese room but the Chinese room as a whole, and for all we have seen thus far there is no reason to deny that the system as a whole can think.
The interrogator is allowed to put questions of any kind to the 2 individuals, and the target can be said to have passed the Turing test if, after an extended session of questioning, the interrogator is unable to determine which room the target is in.
But should we regard passing the Turing test as validation of an entity’s claim to be a thinker? I think not. In fact, many theorists have argued that the Turing test is too weak and that something else is required for true thought.
A central feature of the Chinese room scenario is its appeal to a look-up table. Because of this fact, there is no substantive structure in the way that the system generates the output that it does.
In other words, the look-up table operates along entirely non-compositional lines. Genuine thought, by contrast, is compositional. It is in virtue of this fact that genuine thinkers are able to comprehend sentences that they have not previously encountered. By contrast, a creature that is reliant on a look-up table to generate its behavior will be restricted to the information contained in that table, and will fail to make sense of novel input.
How do natural language symbols acquire their content? My means of conventions. For example, English employs the convention that “monkey” refers to monkeys.
On this view, we do not think in Chinese or Portuguese but in “Mentalese.”
They have persuaded many theorists that even if much distinctively human thought may be encoded in natural language, there must also be a more primitive “language of thought” that is independent of one’s natural language.
Symbols in the language of thought — “ideas,” in the terminology employed by the Empiricists — mean what they do because of what they resemble. In a nutshell, the idea of a monkey refers to monkeys because it resembles monkeys.
There are, however, serious problems with this account. For one thing, resemblance is ubiquitous, and any brain state resembles multiple objects. A second problem is that we can think about all manner of things that do not resemble brain states in any way. In what sense could the mental symbols for
My aim in this chapter has been to introduce the computational theory of thought, and to explore a couple of the leading challenges to it. We have not, however, said much about that the alternatives to it might be. Let us bring this chapter to a close by asking whether there might be other ways in which thought could be realized in a purely physical system.
It is clear we do not have direct and immediate access to the identity of our standing propositional attitudes — our dispositional beliefs, desires, and intentions.
But because these attitudes are unconscious we have no transparent access to them, and there is no reason for the Cartesian — or anyone else for that matter — to think that we will be reliable in self-ascribing them.
Our shared humanity grounds the reliability of a certain range of mental-state ascriptions, but the further we move away from those who are “near and dear” to us, the more uncertain our access to thoughts of others becomes. Even more daunting are the challenges posed by identifying the thoughts of those who lack speech. We will consider such challenges in the next chapter.
Once creatures are able to speak to each other they are able to speak to themselves, and a creature that is able to speak to itself may, in time, acquire the capacity to keep its thoughts to itself.
As the Mad Hatter pointed out to Alice, we do not always mean what we say, nor do we always say what we mean. The interpretation of speech is often tentative and provisional, and we rely on a host of background assumptions about a person’s sincerity, their knowledge of the audience, and the grasp of the meaning of the words that they use in order to infer that they think from what they say.
Where might we look for evidence of thought in a species? We could consider its navigational capacities, for navigational prowess often involves representing the temporal and spatial features of one’s environment in complex and systematic ways. Alternatively, we could look at its tool-making capacities, for turning an object into a tool requires a grasp of its causal properties. 3 other domains: the domain of numbers, the domain of social relations, the domain of mentality.
Social rank plays an important role in many species, and it is vital that an individual knows not only its own place in the social universe but is also able to track the social status of the other members of its group.
The ability to decouple one’s thoughts from one’s immediate environment may be facilitated by (and perhaps even required) the use of symbols.
Aristotle took the capacity for rational thought to be a defining feature of human nature, a feature that distinguishes human being from other animals. However, Aristotle also held that this capacity is not everywhere the same, and that Greek modes of thought were superior to those of other countries.
Pre-literate peoples have little aptitude for logical thought, and that the members of such societies were “uncultivated in following a chain of reasoning which is in the slightest degree abstract.”
Luria concluded that the peasants had impoverished reasoning skills, and drew the more general conclusion that formal schooling is required for the mastery of abstract reasoning.
Two points needed to be kept in mind in considering Luria’s result. Firstly, conscious thought is tiring and effortful, especially when one is considering unfamiliar topics. The fact that the peasants performed so poorly may have had more to do with their levels of motivation. Secondly, in most cultures the only questions that people pose are those to which they do not know the answer.
What does it mean to say that East Asians think holistically and Westerners think analytically? East Asians tend to place more weight on the contextual aspects of a situation whereas Westerners tend to focus on its focal elements; East Asians tend to group objects on the basis of their relations where Westerners tend to group objects in terms o their membership of a common category; East Asians tend to reason on the basis of similarity whereas Westerners tend to reason on the basis of rules.
Although it is generally agreed that the mastery of a natural language has a transformative effect on thought, there is anything but general agreement about whether, how, and to what degree the differences between languages bear on the structure of thought. Some theorists regard the effects of language on thought as deep and profound, while others hold that the differences between languages are negligible as far as thought is concerned. As with many debates, the truth is likely to lie somewhere in the middle.
The perception of light blue and dark blue is more marked in speakers of Russian, a language where the 2 colors have distinct terms. For speakers of English, both colors are described as shades of blue.
The Chinese term for 11 is “ten and one.” Children who are bilingual in Welsh and English perform to a higher degree when they calculate in English, perhaps explained by the fact that number words in Welsh are much longer than they are in English.
The presence of number words in a society can have a profound impact on the mathematical capacities of its members, but words are not the only tools that extend the reach of thought, for human thought is scaffolded in many ways. Thought is scaffolded by culturally transmitted practices, such as the habit of using one’s fingers to enumerate the members of a set, or the practice of remembering a list by imaginatively placing each of its members in a separate room of one’s house. Thought is scaffolded by social institutions, such as schools, scientific communities, and publishing houses. And thought is scaffolded by artefacts of various kinds, such as the sextant, the slide rule, and the smartphone. Thus, even if the basic cognitive capacities of human beings are fundamentally unchanged from one setting to another, the thoughts that are readily available to the members of one society may differ in radical ways from those of another society, for the regions of cognitive space that are accessible to a person depend not only on their basic cognitive capacities but also on the ways in which those capacities are scaffolded, and the scaffolds of thoughts are not everywhere the shame.
There is much that is problematic about this characterization — delusions need not be false, nor need they concern “external reality” — but it does capture the essential feature of delusions: delusional thoughts have become “disconnected from reality.”
When faced with the need to explain an event, someone with a vulnerable self-conception will be strongly inclined to adopt an externalizing explanation (“the fates conspired against me”) and overlook internalizing explanations (“I wasn’t good enough”) that might threaten their already fragile self-esteem.
Delusions can be regarded as “theories” that the subject constructs in order to impose order and meaning on certain kinds of anomalous experiences.
The system predicts what it will feel like when one moves one’s hand. As a result of these predictions, the sensory consequences of one’s own actions are attenuated, and the sensations that they generate are less intense than the sensations that are generated when one’s hand is moved by someone or something else. (This explains why it is difficult to tickle oneself.)
When asked about the basis of his delusional thoughts, John Nash replied that he took his delusional ideas seriously because they “came to me the same way my mathematical ideas did.” The idea that delusions are accompanied by a sense of intuitive obviousness is frequently encountered in the writings of those who suffer from them.
Suppose that I ask you why democracies tend not to wage war against other democracies. If you have not already considered this question, you may need to think about it. What precisely does that involve? Well, if your experience is anything like mine, you put the question to yourself and wait for something to come to mind. Sometimes nothing much comes to mind and the question sits there unanswered; on other occasion, one’s unconscious comes up with something intelligible. Either way, there is no algorithm or recipe that one can consciously follow in order to generate the required thoughts.
On the whole, thinking does not seem to extend much beyond putting questions to oneself and waiting for the unconscious to get around to answering them. The role of consciousness in such cases seems to be restricted to that of a minder charged with ensuring that one’s mind does not wander off topic.
Perception is in this sense passive — it is something that happens to one. By contrast, thought does not seem to be passive in quite the same way. Thought seems to be “spontaneous.”
Although belief-formation may not be under our direct control, we do possess various forms of indirect control over what we believe. For example, we can critically evaluate those ideas that are presented to us as potential candidates for belief. We can take a step back from claims that seem plausible and ask whether the evidence in their favor is as strong as it appears to be. We can probe new aspects of our environment, and thus acquire beliefs about topics of which we had been ignorant.
The advocate of Clifford’s dictum might think that religious beliefs are precisely the kinds of beliefs that we ought not hold but the problematic consequences of Clifford’s dictum go far beyond the realm of religion. Endorsing Clifford’s dictum threatens to undermine our right to hold many of our most cherished beliefs about morality, politics, and philosophy, for these are domains in which it is notoriously difficult to secure consensus, even among reasonable and impartial persons.
The question is whether belief-formation should be governed solely by a concern with truth. What is so great about truth anyway?
Perhaps those of us who have an overly rosy self-conception are better off than those of us who have an accurate self-conception. Certainly there is ample evidence that human beings do in general have an overly positive view of themselves. Most drivers believe that they are better than average drivers; most teachers believe they are better than average teachers; and most people believe that they have a less biased self-conception than other people do. In fact, having an overly positive self-image may even be a universal trait. Rather than assuming that natural selection always privileges the formation of true belief over its alternatives, there may be domains in which self-serving biases in belief-formation have been selected for.
One might also think that giving up on belief in free will or the objectivity of morality has the potential to undermine many of the social and political institutions that are central to our identity as human beings.
Our intellect holds the same position in the world of thought as our body occupies in the expanse of nature.
We are likely to visit only a fraction of the universe, for given our physical limitations vast tracts of space and time are simply inaccessible to us.
The scope of human exploration might be limited, but what about the scope of human thought? Is thought also restricted in its range, or are there no limits on what we can grasp in thought?
Let us know consider 3 aspects of reality which many theorists have claimed to lie beyond our powers of comprehension: the nature of consciousness, the nature of things in themselves, and the nature of God.
What is the relationship between your current brain state and your conscious experience of the coffee? Why is this brain state associated with consciousness, and why is it associated with this particular kind of conscious experience (the taste of coffee) rather than some other kind of conscious experience.
Perhaps we are in roughly the same position vis-a-vis consciousness that mathematicians were in before they discovered zero or physicists were in before they recognized the distinction between velocity and acceleration. Perhaps all it takes is for someone to have a really good idea, and an understanding of the relationship between consciousness and the brain will be within our grasp.
An important philosophical tradition holds that our knowledge of the world is necessarily restricted to the ways in which things appear to us, and that we can never grasp the nature of things as they are in themselves. This doctrine is most famously associated with Kant, who distinguished things as they appear to us (the phenomena) from things as they are in themselves (the noumena). According to Kant, science is limited to the realm of appearances, and it can never provide us with access to what underlie those appearances.
A causal power of an object is the power that it has to modify or affect the objects around it in some way. Kant argues that an object’s causal powers are not intrinsic to it but feature among its relational properties. If that is right, then we cannot have any access to its intrinsic properties, for our only access to the nature of an object is via its causal powers. The interrogation of nature involves “poking” and “prodding” things and then looking to see what effects this has on us. We have no other method for doing science.