We began this meeting with two questions.
First, imagine a dog watches you put a treat in a drawer, leaves the room, returns later, and goes directly to that drawer. What makes it tempting to say that the dog believes or remembers that the treat is there? What changes would make that description less convincing?
Second, suppose three people describe a dog’s state differently:
“The dog believes its owner is outside.”
“It believes the familiar person is outside.”
“It believes someone who will feed it is outside.”
Could evidence make some of these descriptions better than others? Could more than one be equally correct? Or must there always be one uniquely most accurate description of exactly what the dog thinks?
These questions connected the previous meeting’s discussion of thought without language to John Searle’s discussion of animal minds.
Searle distinguishes several things that can easily be run together.
Searle describes consciousness in terms of subjective states of “sentience” and awareness.
We have also been talking about phenomenal or qualitative episodes: experiences such as seeing, hearing, feeling pain, feeling hungry, or being frightened.
I mentioned in class that the word “sentient” is used differently in different philosophical and scientific communities, so do not assume that everyone uses it with exactly the same technical meaning.
The word “intentional” has two importantly different philosophical uses.
In one familiar sense, an intentional action is one done on purpose.
Searle is using an older technical sense. In this sense, an intentional state is a state that has representational content: it is about or directed at some object, event, or way the world might be.
Beliefs, desires, memories, hopes, and many perceptions are intentional in this technical sense.
intentional state = representational mental state
Also note that philosophers use another technical word, “intensional,” spelled with an s rather than a t. That is not merely a spelling variant. We will return to that term later.
Searle also separates having representational states from engaging in thought processes or reasoning.
These three categories can come apart.
Humans appear capable of all three. Searle is confident that many animals are capable of all three as well.
Historically, some philosophers denied animals mentality much more broadly.
Searle briefly discusses Descartes, whose larger philosophical system distinguishes bodies from souls or minds. We will return to those mind/body issues in a few weeks, so we did not pursue them here.
Perhaps animals are conscious, but they do not really have beliefs, other “thoughts” or representational states, or reasoning.
Why think that?
One influential answer is that animals lack language.
Why think language must be required for all thought? Why not instead say that some special contents depend on language, while other thoughts are possible without it?
He then considers several arguments his opponents give for their view.
One argument says:
A creature can believe that P only if it can distinguish believing P from supposing P, merely having a hunch that P, and other nearby attitudes.
When we say things like “believe that P,” we’re using P as shorthand for some sentence, like “Charlotte is not in Florida.” So one kind of “believing that P” is believing that Charlotte is not in Florida.
The argument then adds:
A creature can make those distinctions only if it has the relevant vocabulary and therefore a language.
So a nonlinguistic creature cannot really have beliefs.
Searle rejects the first requirement. Even in our own case, we may sometimes be unable to tell whether a state is exactly a belief, hunch, suspicion, or something nearby. There may even be some indeterminacy about where one category ends and another begins.
That does not show that there is no mental state there at all.
A second argument says:
If a creature genuinely has a thought, there must be a fact of the matter about exactly what that thought’s content is.
But consider a dog whose master is at home. Does the dog believe:
Suppose all three descriptions pick out the same person.
The defender of the argument says that if there is no precise answer about which content the dog has, then we should not attribute belief at all.
Searle rejects that requirement too. He thinks it is perfectly coherent to say that the dog really has a thought while allowing some indeterminacy in how exactly its content should be described. Perhaps some things are clearly part of the content, some clearly are not, and some distinctions are simply left unsettled.
This continues the previous meeting’s examples involving memories and maps. A representation need not take a stand on every possible question.
So uncertainty about whether the dog represents a squirrel, a gray animal, or this squirrel Squiggy does not by itself show that the dog has no representation.
The third argument is more complicated.
It begins with the thought that mental states occur in a network of other thoughts and understanding. If you can think about a gun, for example, you must know or believe various things about guns: that they are physical objects, weapons, and so on.
Likewise, thoughts about squirrels depend on other beliefs or knowledge.
The point of this part of the argument is to show that beliefs and knowledge are primary:
The argument then proceeds roughly like this:
Donald Davidson is one philosopher associated with this kind of argument.
How does Searle respond? He says, let’s grant that belief involves some sensitivity to the possibility of error. Say a dog seems to believe that a squirrel is in a tree. If the dog receives information that the squirrel is no longer there, we expect its behavior eventually to change. If nothing could ever make a difference—if it simply keeps barking regardless of what it learns—we may start to suspect that the behavior is driven by an instinct or rigid disposition rather than a belief.
But from this it does not obviously follow that the dog must possess the abstract concepts BELIEF and TRUTH.
being able to respond appropriately when a representation turns out to be wrong
being able to understand what a BELIEF is, and to explicitly think about one’s beliefs as BELIEFS, and ask whether they are TRUE or FALSE.
Searle accepts something like the first requirement but sees no reason to accept the second.
Near the end of the reading, Searle emphasizes that consciousness and other mental phenomena are part of the “biological natural history of animals.”
His confidence that familiar animals have mental states is therefore not based simply on this argument:
Their behavior looks intelligent, so they probably have minds.
Behavior matters, but Searle treats it mainly as evidence that these creatures have underlying biological mechanisms similar to ours.
We know that our own brains produce consciousness, perception, belief, and reasoning. When other animals have importantly similar nervous systems and behave in relevantly similar ways, that shared causal biology helps support the attribution of mentality.
This sets up an important challenge. What should we say about an animal that behaves in flexible, intelligent, apparently mischievous ways, but whose nervous system and bodily organization are very unlike ours?
That is the issue we turn to next with octopuses.
Make sure you can distinguish:
Does having thoughts require language? Or are the arguments for that requirement really talking about more sophisticated mentality than is needed just to have beliefs and reasoning?