A map is not the same as the territory itself, and a plan or model is always merely a simplification of reality. That is precisely their advantage—while we would never be able to find our way in the mountains or inside the human body without in-depth knowledge, an anatomical model or a hiking map allows us to get an overview in just a few minutes. However, we must also be aware of the aspects that a plan or model omits.
For example, using a cadastral map, we can easily find the desired parcel of land and, based on geological boundaries, its subsoil—but when looking for a restaurant in the city center, this information is more likely to confuse us. What is important to keep in mind before we reduce our mind to a mere processor of inputs and outputs?
The first differences can be found in memory: while a computer has its cache, RAM, and hard drive clearly separated, in humans the boundaries between them are blurred, and information moves gradually from one phase to the next.
Most importantly, however—unlike computers—we do not have separate processes for retrieving, processing, and storing information. Every mental activity involving given memory contents involves all of the listed processes simultaneously: we retrieve information (we call up related concepts from memory and activate related neurons), process it (activations give rise to associations), and store it (any activation of neurons influences their long-term associations, which apparently constitute memory; any retrieval alters memories).
The fundamental differences highlighted by the so-called embodied conception of the mind concern the way information is processed. A computer operates abstractly; software is independent of hardware and the environment. In humans, the opposite is true.
A computer performs theoretical calculations, while we test specific scenarios—which, moreover, are often verified by the body rather than the brain. We’ll perceive a potential partner as more attractive if we have a little adrenaline in our blood: because instead of theoretically pondering whether we’re a good match, we derive the answer from our hormonal response to their presence. When focusing our eyes, unlike a camera, we don’t have to rely on calculations to determine the optimal aperture setting; we simply dilate our pupils until we reach the desired state.
We leave a large number of calculations and operations to our environment, without actually having to perform them ourselves. We automatically associate an office with work and a bed with sleep, without having to arrive at that conclusion through rational thought. Pavlov’s dogs, too, didn’t have to rationally deduce every time the bell rang that they would be fed.
However trivial such automations may seem, they save an enormous amount of computational capacity. When we get into a car, we don’t have to think about whether we wanted to make coffee or turn on the computer—instead, we press the clutch and turn the key without hesitation. Arriving at work automatically puts us in “work mode,” while at a bar we relax on our own.
Similarly, when observing other people or animals, we do not think in theoretical terms (“her eyes are filled with tears, her face is red, her leg is bleeding, ergo, she’s in pain”), but we empathize with the other’s situation as if we were in it ourselves. When we see someone injured, the pain centers in our brain are activated because we do not infer pain, but perceive it; when we see a person raising a glass to their mouth, we know they are drinking because we behave similarly when we drink.
A famous argument against the similarity between human and artificial intelligence is infinite regression. Humans are capable of reflecting on what they are actually thinking about. If a computer were to do this, it would reflect on what it is reflecting on, on what it is reflecting on, on what it is reflecting on… until a programmer comes along and writes an exception into the program. In contrast, humans can escape infinite regression, even though it is not entirely clear how they do so. There are a number of mathematical and psychological perspectives that relativize this argument from one side or the other; however, no one has yet offered a generally acceptable explanation.
It therefore seems reasonable to maintain the belief that human heuristic thinking and computer algorithmic thinking are distinct, just as their electronic and biological foundations are distinct. The computer metaphor can often be useful, but it remains merely a metaphor; a human (at least for now) remains a human, and a computer remains a computer.
