In the last few meetings, we have been asking what kinds of mental states and capacities animals may have, and what evidence could justify attributing those states to them.
Does thought depend on language, or can a creature think without having a language?
To ask that clearly, we first need to separate several different ideas.
Suppose I believe that there is coffee in the kitchen.
We can distinguish:
Belief is only one kind of attitude toward a content. I might instead:
Different people can also have different attitudes toward the same content. One person may believe that it will rain tomorrow, another may disbelieve it, and another may merely hope that it will.
We have been calling states like these representational mental states, because they represent things as being a certain way. Philosophers also call many of them propositional attitudes, because we often report them using expressions such as “believes that…,” “hopes that…,” and “wants that….”
Does having a belief at least require speaking and understanding some language, even if not the language we use to describe the belief?
That is the issue we turned to next.
We distinguished things that occur on the mental side from things that occur on the behavioral side.
Mental capacities include:
Behavior may be communicative without yet amounting to use of a public language.
So we should not collapse these questions:
A creature might satisfy one without satisfying all three.
Also, what matters about a language is not whether it is spoken. Sign languages are genuine languages. The interesting issue is what kinds of structure and expressive capacities a signaling system has.
There is no single uncontroversial line separating languages from every other kind of communication. To some extent, where we reserve the word “language” is terminological.
Still, theorists often look for certain increasingly demanding features.
Some signals primarily express how the signaler feels:
Other signals seem better understood as being about something outside the signaler—for example, a predator, food source, individual, or location.
Some animal signaling systems do seem to have that second capability.
A signal can resemble what it represents (as when a bird imitates the sound of a fire truck), or its connection to its meaning can be biologically hardwired.
More flexible systems allow a creature to associate an arbitrary signal with a meaning. Some animals have demonstrated this capacity.
Some animals can learn multiple signals and combine them in new ways. A creature taught signals corresponding roughly to BANANA, WANT BANANA, and GRAPE might transfer what it has learned and produce a new combination corresponding to WANT GRAPE.
That is impressive. But human languages have much richer forms of combination.
Human words combine into phrases, and those phrases can function as single units inside larger expressions.
Compare:
The italicized subject expressions can become increasingly complex while still occupying the same broad role in the sentence.
Moreover, linguistic structure is not just the visible order of words. Compare:
Pat ridiculed Harry’s theory that Mary was across the street.
and
Pat persuaded Harry’s students that Mary was across the street.
The strings of words look similar, but their grammatical structures are different. For example, Harry’s theory that Mary was across the street behaves as a single phrase in the first sentence in a way that Harry’s students that Mary was across the street does not in the second.
When theorists say that impressive primate or parrot signaling still “falls short of language,” this kind of structured organization is usually what they have in mind.
We distinguished three broad possibilities.
No language, no thought.
The early 1600s philosopher René Descartes (who we’ll encounter again in a few weeks) and the late 1900s philosopher Donald Davidson are examples of philosophers who have defended versions of this idea.
Creatures can already have all the thoughts they are capable of having even without language. Language merely gives them a way to communicate those thoughts to others.
A creature can have some beliefs and other thoughts without language, but acquiring language enables additional concepts, beliefs, and thoughts.
Perhaps some particular concepts depend on linguistic resources. We discussed INFINITY as one possible example. Or perhaps language increases the number and complexity of thoughts a creature can manage even if no single concept marks a dividing line.
The reading gave reasons why, even if language provides especially good evidence of belief, we shouldn’t yet assume that language is required for belief. Nonlinguistic communication, tool use, hunting, food preparation, and socially learned traditions may also provide evidence of thought.
The three possibilities just described concern how thought and language are actually related.
That is different from asking: Which possibilities are we currently in a position to establish?
Someone might think the evidence decisively supports one of the three positions.
I am confident that language is not required for all thought, but I am not yet sure whether language merely expresses preexisting thoughts or genuinely expands what can be thought.
Rejecting one view does not require being equally confident about which remaining alternative is correct.
This is a general lesson about reasoning: when there are more than two possibilities, evidence may rule out some options without uniquely identifying one survivor.
At the end of the meeting we returned to an earlier distinction among:
Some ordinary facts can become indeterminate at very zoomed-in levels of precision. How much exactly does a person weigh? Does the answer include every hair, bit of sweat, food being digested, or flake of skin about to fall away?
Representations can also leave things unspecified. This is another way of being indeterminate.
A memory of a childhood friend might represent:
while simply failing to represent whether there were tiny specks in those eyes, or whether there were more specks in the left eye than in the right eye.
That is different from representing:
There were specks.
and also different from representing:
There were no specks.
The memory may simply take no stand.
Maps work similarly. A map can represent some features of an area while leaving others entirely unrepresented.
This set up the next meeting. If a dog thinks about a squirrel in a tree, must there be one completely precise answer to whether its thought represents:
Or can a genuine thought itself leave some of those distinctions unsettled?
Make sure you can distinguish:
How much thought is possible without language, and what difference does acquiring a language make?