The book is full of relevant, interesting things, so partly as an aid to remembering it, I'm going to summarize what struck me the most from the second half.
Too little emotion
Grandin tells the story from Descartes' Error of a man with frontal lobe damage who seemed completely emotionless (no fear, no happiness, etc.). He showed as having no cognitive disabilities at all on any standard intelligence test, yet the man was impaired. Evidently, he couldn't make any decisions. Grandin says this is because he "couldn't predict the future" -- she believes visceral emotions are a future-predicting mechanism. This makes a little bit of sense when you consider the test that a graduate student finally developed that could detect the impairment that the emotionless brain-damaged man had. It was a "gambling test." In the test, the subject starts out with some play money and four decks of cards. He is told that every card he turns over will win money, but some of the cards in the decks will lose money also. The goal is to keep the original money and win more if possible. It turns out that decks A and B have big wins but also big losses, and C and D have smaller wins and smaller losses. If you do the math out, drawing only from decks A and B will cause you to lose all your money, but drawing from C and D will result in coming out ahead. A normal player will play from A and B for a while because the payoffs are high, but after their money has dwindled a certain amount the normal player will get a "bad feeling" from those decks and move to C and D. The brain-damaged man, however, never got the "bad feeling." He just kept drawing from A and B and going deeper and deeper into debt. So apparently, the brain-damaged man couldn't make any decisions because he couldn't attach an emotional preference to any particular choice -- he would just deliberate forever. The researcher working with him offered two days for an appointment, and the man went on for a half an hour debating with himself the pros and cons of each day, never coming to a conclusion until the researcher chose a day for him. Grandin says, "Emotion is so important that if you had to choose between having an intact emotion system in the brain and having an intact cognitive system, the right choice would be emotion" (p. 198).
Animals and Language
Grandin discussed animal intelligence a certain amount -- touching on the "Shaping of Hooks in New Caledonian Crows" findings and the Alex the Gray Parrot research. However, she also mentioned two interesting things I'd never heard of.
The first was research into prairie dogs' distress calls. Apparently, prairie dogs have a very sophisticated aural communication method -- one including interchangeable nouns, verbs, and adjectives -- by which one prairie dog can alert all of the others in his community of a predator. The distress calls can say what kind of predator it is, what it is doing (e.g. whether it is hovering, moving farther away, or attacking) and in some cases can even identify which individual predator (i.e., a certain particular coyote they are familiar with) is nearby. Researchers have analyzed the sound spectra of the calls and experimented with mocked-up objects and found that the prairie dogs can even construct ad hoc descriptions that the other prairie dogs, judging by their reactions, understand. These were very surprising findings, and the question that naturally arose was, why did prairie dogs (of all species) develop such a sophisticated language, which we thought only dolphins and much higher mammals capable of? The most likely answer is that they are such vulnerable prey species, they *had* to for self-defense. Prairie dogs live in the same holes for hundreds of years, and their predator list is extremely long. This is interesting when you consider (and the following is my speculation) that proto-humans, millions of years ago, may have been in a similar fix and might have reacted to the same pressure in developing language. If you imagine some naked pre-humans on a savannah, without any tools yet, they were probably fairly vulnerable creatures. In fact, this is one thing that led Elaine Morgan to propose the Aquatic Ape Hypothesis, the argument that there is no way pre-humans -- left on an open savannah with the loss of their tree habitat -- could possibly have survived to evolve into humans. So maybe humans developed language because they spent evolutionary time as prey animals and needed communication to survive.
The second interesting thing Grandin described was research into humans who don't have language. Apparently there are a certain number of deaf people out there who were never taught sign language and so just never learned any language at all. Grandin cited the work of one researcher who taught sign language to a 27 year old deaf-mute, language-less Mexican immigrant -- A Man Without Words by Susan Schaller. Apparently he grasped the concept of language in about six days of work with his tutor, as a sudden breakthrough, although learning to actually use the language took much longer. What's especially interesting is the insight this provides into what the man understood before he had language, and what he didn't understand until after he had it. "For instance, he didn't have the concepts of real and fake. He just knew that some Green Cards worked to keep the green men from taking you back to Mexico, and some Green Cards didn't. He didn't know why. / He also didn't have just and unjust. It's not that he didn't have morals or a conscience. ... [here, a vignette showing he could act on a natural sense of rightness] ... [but] He had no idea no idea whether it was just or unjust for the green men to catch him and take him back to Mexico; he just knew that's what the green men did, so he needed to stay away from the green men. He was trying to understand the rules, without realizing there were principles behind the rules. / Ildefonso was an innocent. He didn't see all the good and bad that people do, and he didn't know there could be good and bad rules, either. After he learned language, he was sad to learn of the terrible things people do. Animals are innocents, too. Even when animals are treated badly by humans, or see other animals treated badly by humans, they don't seem to develop the abstract categories of just and unjust. ... Animals live much closer to the plain facts of the situation. / But the important thing to realize is that Ildefonso's innocence was not the same thing as being stupid, or unable to think. Ildefonso wasn't stupid, and he functioned as a person of ... even above-average intelligence, given that he had been able to immigrate to a foreign country, find work, and manage his life while struggling with a huge disability." (pp. 257 - 258).
Grandin's view is that animals are smarter than we think, and you don't have to have language for intelligence, either. Animals, she thinks, mainly just have (a) less cultural evolution (where later generations benefit from innovations made by earlier ones), in addition to (and perhaps partly because of) (b) a different *kind* of intelligence than humans have -- specific rather than general.
Hyper-specificity and Autism
Grandin believes that animals and autistic people are hyper-specialists, where normal people are generalists. For instance, an autistic child was taught to butter toast after long effort, and when the adults were done celebrating they handed the child some peanut butter to see if he could spread that too, and he couldn't. The adults had to start all over teaching him how to spread *peanut* butter, which to the autistic child was a completely different thing. Now, the autistic child definitely did generalize to a certain extent -- he learned to butter multiple possible pieces of toast, with one of any similar possible knives, using one of an unlimited possible number of sticks of butter. But his generalization was very narrow. Animals are like this too -- for example, it is considered a good pratice to always mount a horse from the same side, either the left or the right, because if you tried to mount a horse from the side he isn't used to, he might panic because it would seem too unfamiliar. As another example, if an ant retraces an earlier path and comes to a landmark he previously passed, he will go past the landmark a little ways and then turn around to look back at it, in order to see how the landmark looked when he was approaching it for the first time. He needs to do this becase it looks different enough from the two angles to appear to be two completely different things. Grandin says she does this too -- if she's driving and passes something from the opposite side as before, for example a red barn, she will stop and look back at it to be sure it's the right one, because to an autistic person, a red barn looks like a *different barn* depending on what side you are looking at it from.
So Grandin believes that animals and autistic people do not integrate perceptions into concepts quite as readily as normal people do. This is one reason some are savants that can do things like draw an object from memory using perfect perspective. A normal person would translate the visual details into an object -- "a building" -- and then will try to draw the concept. But an autistic drawing savant has privileged access to the pre-conceptual sensory data, and can draw directly from that. This is one reason normal humans have "inattentional blindness" -- because in order to conceptualize, they have to throw out a lot of sensory data and fill in the blanks with what they are expecting. This means normal people miss a lot of details that autistic people see, and even though autistic people can have problems from their difficulties with generalizing, they can also find uses for their strengths, for example they can do well as quality assurance personnel, proofreading or spotting flaws in products.
Grandin's points about fear are pretty fascinating. First, nervousness/high-strung-ness seems to be related to both intelligence and curiosity, and inversely related to aggression. In an early chapter, Grandin explains how fearfulness and curiosity are related. The fearless cow breeds, besides being unflappable, don't seem to care what odd things might be placed in their pasture, but the nervous ones that fright easily will also, almost ironically, be the most aggressive in approaching a strange object and exploring it. The connection is that the more you have to fear from a potentially dangerous thing, the more it pays for you to try to find out something about it (e.g., discover if it is really harmful or not). Later, Grandin observes that high intelligence probably comes from high curiosity and goes with the nervous temperament that way; more exploration simply leads to greater learning. As anecdotal support, she points out that Arabian horses (the most flightly ones) are also often recognized as the most intelligent breed. But also interestingly, it seems that the animals that have had the most fear bred out of them can also be the most aggressive. Fear is a check on aggression, and so calm animals (unless they are ones that specifically have had *both* fear and aggression bred out) can also be the most dangerous -- they fear no consequences of their actions. The effect of handling seems to be directly related to temperament, too: rough handling tends to destroy sensitive nervous-temperament animals, whereas calm, tough animals can habituate to it, even though it may never be the *best* handling they could receive.
Some fears are learned from watching others react fearfully. On pp. 209 - 210, Grandin describes an experiment in which lab-raised monkeys had to observe other monkeys freaking out at the sight of a snake in order to learn to fear snakes themselves. Show a lab-raised monkey a snake for the first time, and he's not particularly concerned. Show a wild-reared snake, and he is afraid -- viscerally, thoroughly afraid. The question researchers asked was whether snake fear was genetic or learned, and the answer is that it's *both*. After showing the snake to the lab monkey and eliciting no response, you put the lab monkey and the wild monkey together and show the snake to both of them. The wild monkey acts afraid, the lab monkey observes it, and from *then on* the lab monkey will be as afraid of snakes as the wild monkey was. The lab monkey imprints the reaction he observes. They found that a wild monkey that just gets a little anxious imprints that reaction on the lab monkey, and a wild monkey that gets petrified imprints that stronger reaction on the lab monkey, so it's not only the reaction or lack of it, but also the degree that matters. Furthermore, putting two lab monkeys together so that each of them can see the other monkey not acting afraid at the sight of the snake makes them both immune to future imprints, even if sometime in the future they are put in the company of wild monkeys freaking out at snakes. So, the *potential* to be afraid of snakes is inherent in monkeys, but not the fear itself. By contrast, a flower was impossible to make a monkey learn to be scared at, no matter what they tried. So, some fears can be learned through indirect transmission, i.e. without direct experience with the dangerous thing, but *only* if the genetic groundwork for that fear already exists.
Fears also can creep. If you experience a fear reaction, and you don't know what the reaction is to, you stand a chance of developing new, secondary fears -- "downstream" fears -- of other things that happen to be present at the time of the reaction. For example, if you are a war veteran who had a terrifying time on the front lines, and as a result you have panic attacks at the sound of cars backfiring, but you don't consciously remember what is causing the panic attacks, then you stand a chance of developing a fear of, say, walking down the street, or reading a newspaper, or riding an elevator, etc., if you hear a car backfire and experience the fear reaction while you are doing any of those things.
Unfortunately, though, fears, once learned, never seem to go away. They are like "riding a bicycle" -- you don't forget. You can learn to tolerate them or mitigate them (by minimizing your exposure to the stimulus, for example), but you never really get rid of them, the same way you will never forget where you were on September 11.
Finally, Grandin cited research that showed that small mammals' fear at predators is not engaged when the predators are sensed using their distance olfactory system, only with their close-up olfactory system (when the predator is practically almost upon them). At first blush this sounds ridiculous because a typical fear response is to freeze up, and this seems to mean that animals with this trait would never run from a predator and would end up as lunch, but it turns out this works because many predator animals are territorial, so smelling where the predator has *been* can be a good indicator of where it *will be*. Also, Grandin speculates that most mammals other than humans have good enough senses of smell that they could be constantly aware of the distant location of *some* predator, so if this distant perception activated their fear circuits they would never come out of the grip of anxiety long enough to eat and mate. The result is that interestingly, a prey animal will become downright fearless when a predator is just out of range, and/or if the predator is tethered or restrained and the prey animal knows it. This reminded me of people who like to procrastinate until the last minute. They don't feel the pressure of the deadline until it's almost upon them -- their fear centers aren't activated.
Physiological correlations to temperament
A minor theme in Grandin's book is the connection between mind and body. (She says, "Don't ever let anyone convince you that Rottweiler or pit bull aggression is a 'myth.' It's not. Temperament and appearance are connected. [p. 237, emphasis in original]) She correlates fine bone structure (e.g., in Arabian horses) to the nervous temperament, and big bone structure (e.g., Clydesdales) to a calm temperament. She notes that the hair whorl (the place on a horse's forehead where the hair twirls around and radiates from a single spot) can show temperament based on its position -- a high whorl means a high-strung, intelligent horse, and a low whorl means a calm one. Another minor theme throughout her book is that animals with partial albinism (white spots in fur, pale eyes) can be somewhat whacked out. In particular, they can be "more inquisitive" and "less wary" -- she mentions several examples, like certain wild otters that allowed themselves to be photographed close up and a wild black bear named "White Heart" (for the patch of white fur on her chest) who was the friendliest one to the explorer who named her, and also the first one to get shot by hunters because she had the least fear of humans. Mating blue-eyed huskies can cause problem offspring, and a study at the University of Utah concluded that researchers should not be using albino animals in their research, because albino animals are not normal. Grandin points out that it's the white *skin* that is correlated to problems, not white fur, and the reason is that melanin (the chemical which gives skin its color) is also found in the midbrain, "where it may have a protective effect." She says that dalmations and "paint horses" (black and white ones) can be "plain crazy," especially if they have blue eyes. (Note: "white" people do still have melanin in their skin, only relatively smaller amounts, so this phenomenon doesn't apply to white people per se.)