Knowing dogs

kw: book reviews, nonfiction, animals, dogs, science journalism

I am not much of a dog person. The three dogs I owned all lived less than a year, each dying when struck by a car. By the time I was fifteen I was done with dogs. I had better luck with cats, keeping one for 18 years, but that is a subject for a different article.

Can you tell what this dog is feeling? Wouldn't you say he has guilt written all over him? I find it interesting that, seeing a furtive look on a friend, we might say he has guilt written all over his face, but of a dog, we generalize to the whole body. Dogs speak with their whole body, while human "body language" is largely confined to our faces.

Until recently, no scientist would agree that this dog is feeling guilt. That was considered unwarranted anthropomorphism. No animal was allowed to have feelings in any way mirroring human emotions, or thoughts of any sort. The paradigm in Animal Science courses of the 1960s was that animals didn't think, had only the most primitive emotions (left unnamed), and didn't genuinely feel pain (which allowed experimenters to inflict agony without feeling guilt themselves).

Jane Goodall changed all that, by naming the chimps she observed, and writing about them almost as though they were human. After terrific resistance, the scientific community began to shift. Now animal scientist Alexandra Horowitz can write Inside of a Dog: What Dogs See, Smell, and Know, and escape excommunication from the scientific community. As she tells us, dogs do see, though differently; they do smell, and much better than we; and they certainly know things, but having much different interests they also know things we don't know, just as we know much that they won't ever know. The book's illustrations, like that above, are by the author.

It starts with size. Attach a video camera to your knee or even lower and go about your business while it is running. This is better than attaching it to your dog, who just may object to being "critter-cammed". How does the world look from about one foot above the ground? Even more so, how does it smell? Our noses are four to five feet above ground level, where the scents that waft upwards from ground-bound "things" are much diluted. A dog's nose is naturally carried less than a foot high, and can easily be lowered right to ground level. So even if a dog's eyes were identical to ours, and his nose identical to ours (a tragic thought), it is like real estate: Location, Location, Location!

But a dog's eyes are different from ours. They do see most colors, but not red (neither does that bull in the field see red; a red flag or cape looks black to him, making him think he is being challenged by a rival bull). A dog's vision is from greenish-yellow to violet. In addition, most dogs see horizontal motion much more keenly than we, the better to detect that rabbit starting to turn aside. Yet for many of the things we see well, dogs' eyes are less efficient. A dog's vision is used two ways: firstly, to spot things worth smelling or tasting, and secondly, to observe us and know us. They attend to us most persistently.

(I have heard preachers say that only humans have religion, because "you never see a dog making an idol to pray to." They don't have to. They live among their deities! Think about it. We feed and shelter them, heal their illnesses, and freely exercise the power of life or death with them. We are their gods. If you lived in your god's house, and could see your god, would you not be very, very attentive?!?)

But the eyes are secondary to the nose, the nose! Most dogs are a million times better at smelling than you or I. Not only do they have 10,000 times as many smell cells; they have anatomical flaps and spirals that do a much better job of getting those smells to the cells. Where "seeing is believing" for a person, dogs only believe what they smell. A person your size, disguised as you, might fool your dog for a moment, but as soon as the air shifts the imposter's scent the dog's way, he'll know right away! Thus, if you are walking your dog, don't drag him away from every smell. That is like making a friend you are walking with go blindfolded.

Dogs have co-evolved with humans for 15,000 years or longer, perhaps much longer. The original wolves who worked their way into human camps and, eventually, homes, were those who could tolerate human nearness and contact; those who were best able to respond socially to us social apes. Experiments with Siberian foxes, begun two generations ago, showed how quickly one could turn wild foxes into a doglike domestic animal, just by selecting those which were least fearful of humans for breeding, and the same with their offspring. In the time it takes a human baby to grow up, about fifteen fox generations, the experimenters had a bunch of foxes that liked human contact, barked instead of yipped, had softer ears and more curly tails, and higher foreheads. All these characteristics seem to come as a package, genetically. There's a PBS video on this; it is uncanny seeing foxes that look like dogs and act like dogs.

But what do dogs know? Can they know as we know? Have you ever watched a sleeping dog twitch during a dream? Maybe she yelps quietly and her legs kick. She is playing with a friend, or chasing something, in her dream. Now think about your own dreams. Do you speak and hear words in your dreams? I do. Do you ever smell anything in a dream? I never have. It seems dogs always do, as much as can be told from watching them and noting how they snuffle during a dream as during waking. We tend to think in a combination of sights and words. We know what we can describe. Dogs thinking has to be different, based on sounds and bodily feelings (for they do love the tactile sense). I, at least, am convinced that any animal (not just vertebrates) thinks and feels things; that is what brains are for. Dogs happen to be more compatible with humans than any other animal, partly because there is quite an overlap between our ways of feeling and thinking.

The book's title comes from a saying by Groucho Marx (or one of his writers): "Outside of a dog, a book is man's best friend. Inside of a dog, it's too dark to read." Dr. Horowitz spent more than a year observing, videoing, analyzing, and learning a bit more how to understand that postural language, Dog. After reading her book, inside of a dog, it is getting easier to read.

From eyeglasses to cybernetic eyes, and beyond

kw: book reviews, nonfiction, science journalism, technology

People's ideas of robots is most recently influenced by the Wil Smith film, I, Robot. Though the title is from Isaac Asimov, the theme is due more to Jack Williamson, who foresaw robots becoming our nannies, to a pathological extreme, in his story "With Folded Hands". Really, though, where is robo-technology going?

"The future promises a wealth of humanlike machines and machinelike humans," declare Gregory Benford and Elisabeth Malartre in Beyond Human: Living with Robots and Cyborgs. The viewpoint is that robotic and cyborganic technologies are tools, with more "smarts" built into them, perhaps, but no more innately good or evil than are eyeglasses, hearing aids, pacemakers, or titanium knee replacement joints.

The authors survey the spectrum, beginning with "Man Plus", which indeed includes eyeglasses and not just hearing aids but lens- and cochlear implants, and trends inevitably toward, perhaps, the brain getting either a "body transplant" or a mechanical body, à la Robocop. Is there a line somewhere beyond which the person ceases to exist? Or is it somewhere between Robocop and the "downloaded personality", in which the brain also is an artifact? Or is there no line? They make their position clear that a bodiless brain/mind has no meaning; our bodily functions and emotions are essential to our mental functioning, even though at times they may seem to hinder it.

They then survey the field of "Robots Plus", the spectrum from industrial welding robots, in use by the thousands, to humanoid, or rather "mammaloid", types, including Honda's Asimo and various robotic pet dogs. There is quite a bit of sociological musing here, with the conclusion that while it is necessary that some robots appear to be human, there is the danger of people imputing too much humanity where it isn't deserved. We're all to willing to lean 'way over backward to "help" a mechanism seem to pass the Turing Test.

I am reminded of a simple computer program I wrote, based on a story I read. It answers questions with Yes or No, implying a 20-questions game. After the fifth Yes answer, it asks "Is that it?" If the user enters Yes, it goes on to "Start Over?" People who play it express amazement at the database that must underlie its ability to guess their hidden thought, when they actually supply nearly everything themselves. One version of the game simply answers Yes every time the last letter is an E; others use a slightly more complex calculation (number of vowels, perhaps). But those who use it find themselves thinking it has great abilities.

The third part of Beyond Human explores the middle ground, a meeting place as it were between mechanism and human. At this point I began to do some calculations. All who viewed I, Robot were impressed with the power and speed of the malicious robots, even though it was found that most of their brain power was at the other end of a radio link. Just how quick and strong can a robot be?

The heavy welding arms in factories are known to be dangerous. They don't sense their surroundings, and a few techs have been injured or killed when one swept an arm through a space that is ordinarily free of human obstructions. But they are attached to and driven by heavy machinery "below deck", and the entire mechanism is the size of an auto. What are the physical limits of an autonomous robot?

A human athlete weighing 75 kg has a basal metabolism of about 75 watts, or 1550 Kcal/day (the nutrition "calorie" is a Kilocalorie). During heavy stress, such as a bicycle race, or lumberjacking, a trained person can exert 700 kcal/hr (about 800 watts, or one horsepower) for extended periods. An athlete in training can carbo-load before a major event such as a marathon, to gain a "quick reserve" of 8,000 kcal or more. The athlete's fat reserves total another 30,000 to 50,000 kcal. The entire energy system uses about 60% of one's body weight, or 45 kg for this athlete. To put these into electrical terms, 8,000 kcal = 9,300 watt-hours, and a 40,000 kcal reserve equals 46,500 w-h.

The two best battery technologies available, Li-ion and Li-thio-chloride (LTC), can store 128 and 700 w-h per kg, respectively, though the latter cannot sustain rapid discharge. An experimental technology by Nanoexa exhibits storage density of 3,000 w-h per kg, using nanotechnology to increase the efficiency. This last implies that a 3 kg battery could store the energy found in a fit person's fat reserves. If Li-ion is needed for faster discharge, however, the mass requirement is 360 kg. Even the "carbo-load" reserve requires 75 kg of Li-ion batteries. More to the point, just keeping a human body alive for a day uses the energy found in 15 kg of Li-ion batteries.

Perhaps we need to think of other types of animals. A 75-kg American Alligator has only 1/25 the resting energy need of a human (about 3 watts). That is why they can stay underwater for so long, while we can barely stand to hold our breath for one minute. Having a small heart, and modest short-term reserves of glycogen, Alligators work best when they can work in short spurts, so they are ambush predators. Not for them the Cheetah's run or the Bear's charge. A mechanical alligator could do short spurts of energetic work, but would need a powered umbilical for sustained effort, an option people don't have.

But the power-to-weight ratios of smaller bodies are even more unfavorable. Much is made of "robotic insects" for surveillance. A crawler and wall-climber might work out, but a flying robo-bug simply can't carry enough energy to stay aloft more than a second or two. It makes more sense to use an ultra-small "critter cam" attached to a Cicada or large fly.

Finally, the authors bring out a point that few science fiction writers ever consider: The marketplace will decide what sells. The technologies available forty years from now will be those that pass the market and utility tests. Robots for sale in 2048 will be robots people like, and enhanced humans cannot be too obvious, or they will suffer from the "green monkey" syndrome and be shunned.

First, assume a spherical cow of uniform composition...

kw: book reviews, nonfiction, collections, science journalism

OK, I assume you know the joke about asking a physicist to help improve the yield of dairy cows. That's how most folks look at physics, isn't it? Physics is math on steroids, right? Here's another one:

Are you an engineer or a mathematician? You enter a room. A sign says, "Boil the Water." You see a mug of water on a table just to the left of a microwave oven. OK, no prob; just put the mug in the microwave, and heat for a couple minutes.

Another sign says, "Go through the next door." When you do, you see a sign, "Boil the Water." You see a mug of water on a table just to the right of a microwave oven.

What do you do? Would you put the mug in the microwave, and so forth? If so, you're an engineer. A mathematician would move the mug to the left of the microwave, and state, "I already solved the rest."

What is Physics, really? Jennifer Ouellette's book will give you a good beginning, of seeing how physics has progressed during its 500-year history. Of course, "physics" wasn't its name in the 1500s. Instead, it was the major element of "Natural Philosophy". The 38 re-edited columns in Black Bodies and Quantum Cats: Tales from the Annals of Physics present highlights of the past half-millenium from a very human view.

The author is editor of American Physical Society News (APS News), and her monthly columns have been used by physicists worldwide to help their "physics-phobic" friends understand a little of what they do. Physics is not just math and esoteric experiments. It is a most human endeavor, and the stories Ms Ouellette tells are very human. Her love of history and fascination with the way science really happens come through loud and strong.

I am going to take a long sidetrack here. Nobody can be an expert in everything, and the author's blind spot seems to be optics. In particular, when discussing the compound microscope, and then telescope optics, she gets something very crucial completely backwards. I'll explain.

Simple, single-lens microscopes worked better than compound microscopes until the invention of the achromatic lens, which uses two lenses of different types of glass to cancel out most of the color aberration that is present in any single lens.

A single lens is used as an adjunct to your eye's lens to let you look at something really close. Most younger people's eyes work best at distances of ten inches or greater. We use a ten-inch distance as a benchmark. A lens with a 1-inch focus allows you to look at something one inch from your eye, so it looks ten times larger. The smallest lenses made by Leeuwenhoek in the 1680s had a focus of 1/30th inch, so they "magnified" by 300x. But you had to hold the lens so it nearly touched your eye. That is why his microscopes are built the way they are.

The compound microscope uses a kind of image relay to achieve large magnifications without sticking something right into your eye. If you have a large magnifier, try this. Instead of holding it close to your eye, hold it farther away, at arm's length, but a few inches from something. If you hold it close to the object, you see the object a little magnified. As you move the lens farther, the image you see gets bigger, then starts to be really distorted. Farther yet, and you see an enlarged image again, but upside-down. It may be hard to focus on also, because it is closer to you than the lens is. That is called a "real image", because if you put a piece of paper there, the image will be visible on it.

A compound microscope works by making an enlarged real image, then magnifying that with an eyepiece lens. For example, an ordinary school microscope might have a 10x objective, and a 10x eyepiece. The objective is made so that when it is 16mm (0.63 inch) from an object (the "slide" on the microscope stage), it produces a real image 160mm (6.3 inches) farther up the tube. Another 25mm (1 inch) farther up, the eyepiece, with a focal length of 25mm, works with the lens of your eye to focus the image on your retina, and it is magnified 100 times, compared to the slide.

While you can make a rudimentary compound microscope with two simple lenses, there will be strong rainbow effects, because the color aberrations of the second lens will multiply those of the first. So, special glasses are used to make lenses that don't have much color aberration, so you can see a clear image.

Now, in Chapter 3, Ms Ouellette states that the lenses in a compound microscope are "complementary", "one convex, the other concave." She states elsewhere that the concave lens (which spreads light, doesn't focus it) is the objective, while the convex lens (which does focus light) is the eyepiece. I have seen one of these historical microscopes, and I can assure you, both lenses are convex. What is going on here? Where did she get this idea?

It may be from an imperfect understanding of achromatic lenses. These do have a complementary pair of convex and concave, but the convex is the stronger. We want these lenses to focus, after all. But I think it might have been from partial knowledge of Galieo's telescope, and of common nautical "spyglasses" of the 19th century and earlier. These do use a convex and concave lens, one at each end of the tube. However, the convex lens is the objective, and the concave lens is the eyepiece. How this works is subtle.

The long-focus convex lens at the "business end" of a telescope has the job of producing an real image near the observer's eye. The eyepiece then has the job of relaying this image into the eye. Galileo and others of his time used a concave lens, which is easy to use. Concave lenses don't focus light by themselves. However, when light is coming to a focus because of a nearby convex lens, a concave lens can shift the focus, or even reverse it, so to speak.

A concave lens has a "negative focal length". If you produce a real image using a convex lens, then put a concave lens at just the right distance from that image, but intercepting the light before it gets there, it will just counteract the focusing effect, so the light seems to be coming from an object much farther away, but of a different size than the original object being imaged. They, when you place your eye near the concave lens, this nearly-parallel light will be focused by your eye's lens onto the retina and you'll see a magnified image. If you move the concave lens a little closer to the convex one, the size of the image will decrease a little, and you'll need to focus your eye as though the object were closer, but it is like a zoom lens.

This makes spyglasses easy to use. There is one setting where you get maximum magnification, and your eye's lens is relaxed. There are other settings with a little less magnification, as long as you can accommodate to viewing an image that seems to be closer. Of course, if you push the eyepiece in too far, you can't focus at all.

If this is so easy to use, what is the drawback? The geometry of the arrangement means you have a very narrow field of view. Looking through a spyglass, you see a small circle with a magnified image in it. But when you look through most modern telescopes, you see a much larger circle. Sometimes, you even have to look around to see the edges of the view. Why?

This is because they use convex lenses as eyepieces. A "positive" eyepiece stands farther from the real image produced by the objective, and relays it to your eye. Because a convex lens is a focuser, it can gather light over a larger angle and put it into your eye, so the view is much wider. There is a further, really big difference. A spyglass has its view right-side up, while a modern telescope has its view upside-down. There are simple ways to turn the image back over if you need to (binoculars use internal mirrors or reflecting prisms). Finally, though, the focusing range is more narrow, compared to a spyglass eyepiece. That is why most spyglasses just have a slide tube holding the eyepiece, but most telescopes use a mechanical focuser.

So, whether it is a microscope or telescope, an objective lens produces a real image of an object, and an eyepiece accommodates that image to your eye, further magnifying it in the process.

OK, back to the book. I really like the author's approach. She uses everyday analogies and popular culture (Addams Family Values, or Back to the Future...) to associate physics concepts to more familiar things. So, let's think about this. What do you do when you get curious about something?

Did you ever see the rainbow in the spray of a sprinkler on a sunny day? If you look closely, you can see that as a droplet of water moves, the color it sends to you changes. Perhaps close to the sprinkler head, it is silvery, but not very bright. Then as the drop moves up, it'll brighten to blue, green, yellow, red, then back to a dimmer, silvery color. Perhaps you'll see it fall back through the colors in reverse order before it hits the grass. If you want to, you can hold a drop of water in a tiny wire loop indoors, and shine a flashlight on it, to see what angles give you what colors. You're doing physics!

The big, theoretical formulations by Einstein, Dirac, and others, are one thing. Yes, there can be some really heavy math involved. But the foundations of physics rest in simple things like rolling a ball on a tilted table with a stopwatch in your hand. Ms Ouellette is to be thanked for bringing physics a bit closer to the everyday world.