Tuesday, June 26, 2012

The Forever Man - Part 1 - To Live is to Serve - Pages 3 and 4

It was an old hatred, although how old he could not be sure. He knew he had been young -- not younger, but young -- and new to his place. His travails at the desert's heart had come later. Before now, and before that, he had been sent over the wall.


Whoever had built the wall had built it well, to seal the enclave off from the hazards of the outside world. Not only was it too high for bugs and herding animals to scale, and too strong to be destroyed, it had neither gates nor doors nor apertures, no weak points in its sheer surface to allow an external enemy to gain access. Unfortunately, those master builders, whoever they had been, had not foreseen everything. Years earlier, before Petros' time, or so he was told, the machines which fed the enclave had failed. Wonderful devices they had been, he was told, able to create food and water from the light of the sun itself. But, he was told, the machines had failed, either through age or neglect, and the only food for the Pulmeks came from outside the walls.

He had been told many, many things by the old and by the dead, and most of it he dismissed as hazy memory, hopeful myth or some combination. But this had the ring of truth, if only because the machines still existed, still sat in the base of the great tower at the enclave's heart, silent and immobile. And, of course, because of excursions like this: the periodic sorties over the wall, into the realms of the enemy.

Petros was newly a Lefent, recently mere Agnant, and he had yet to be given a group of his own to command. For now, he was meant to watch and to learn, to see how Zdeti and Jary and Sten and Kuoni and Cene -- superiors all -- guided and guarded the Pulmeks, their enclave, and the hospice rooms of the dead. Even now, as he prepared at the top of the outer wall to drop down below, guided by his skinsuit's power and his own skill, he was not in command. Command was held by Genal Cene, friendly and genial and -- in Petros' opinion -- simply too weak to hold such a vital position. Commer Zdeti trusted her judgement, for some reason he couldn't begin to fathom, so the story went. And yet she was here, commanding a group far below her stature, readying them to foray out for food and water.

There were three others in the group besides Cene and himself, all Agnants. Swith he didn't know particularly well; she had been on an alternate rotation from his, thus he had only seen her from afar, during raids and other battles. She had been impressive, to be sure, but other than that, he knew little.

By contrast, Adir, standing off to his right and checking the straps on his pack, was an old -- well, what would you call him? Not a friend, as Adir seemed to have no friends at all, spending what little free time he had on his own, designing and planning improvements to the enclave's buildings. Not rival, for, as he had told Petros on numerous occasions, he really didn't care what others thought of him, not even Zdeti himself. Acquaintance, perhaps, or comrade. Certainly reliable, and a redoubtable fighter.

A frown crept over Petros' face. Now that he thought about it, this was quite an unusual group to send out foraging. He had never heard Swith described as possessing great ability as a hunter or gatherer, nor Adir, and certainly not Cene or himself. And the third Agnant, Jian, was more a thinker and strategist than someone who could venture past the walls and thrive.

Something else was going on here. He knew that Cene would never tell him what; was fairly certain that Cene would never reveal the reason she was leading rather than a Lefent like himself. But the others, assuming one of them had some clue, might be worth pressing for information. Jian, possibly. He seemed the weakest, in terms of personality, the one most likely to wither under sustained pressure. He was also, Petros knew from experience, the most intuitive and intelligent of the group, which suggested that he was likely to know almost as much as Genal Cene.

He brought himself up short, realizing that everyone else was looking at him expectantly. He cleared his throat. "Apologies, Genal. I was in thought."

Cene rolled her eyes at him and grinned. Petros considered both insincere, and clearly forced. "To repeat," Cene said, "we will proceed under Lefent Petros' command -- assuming he can keep his attention on our mission. It will be his responsibility to keep us safe in the deserts, and return us here with as much food and water as we can carry." She gestured to the canteens and packs they all carried. "In those, of course. Keep your arms free. It's never quite as safe out there as you might think. Whenever you believe that you can load yourself up and drop your guard -- that's when you're killed."

That shouldn't have been a surprise, Petros thought, as he nodded assent. What better way for Cene to do what she had been ordered to, in secret, than to keep him occupied with a group of unskilled Agnants, put their safety and success in his hands? He would want to prove himself, the Agnants would want to push him and see what he was made of, and the deserts themselves would throw challenge after challenge their way. It was a simple way to keep everyone from looking directly at Cene, distract their attention away from her actions.

He needed a better plan. He walked towards the edge of the wall, the Agnants and Cene trailing behind. He needed a way to get Cene to reveal herself, to make a mistake that would show him what her actual intentions were. He had no serious doubt as to whether her plans came from Zdeti; of course, they must. However, newly-promoted, he was still ambitious, still looking for more authority and power. What better way to impress the Commer than to outshine his hand-picked Genal, to prove himself capable of her position and responsibilities? But, to do it, he needed a plan. And now was not the time to think of one.

Raising his hand over his head, Petros paused at the edge of the wall, one foot braced on its lip. He dropped his hand sharply, and stepped out over the edge.

Friday, June 22, 2012

Practical Philosophy - Chapter 1 - Logic and Language - Page 4

Are there reasons to believe that we can improve the way we think? One obvious, if rather trite, reason is to note that, if we can't, then education is really pointless. If you can't think any better than you already do, then why every learn anything? Your cognitive abilities won't really advance; at best, you'll be refocusing the cognitive resources you already started with. Which only counts as improvement if shuffling a deck counts as dealing a winning hand.

As I said, though, this is a trite point. For one, it may be that learning really is pointless. For two, it's plausible that learning doesn't extend our cognitive resources so much as it activates ones which were already present, but dormant. Consider the use of language, for example (and, yes, I am eventually going to get back to language as my primary topic). It seems to be widely believed that humans acquire language some time after birth through a process of exposure. Parents, siblings and others talk to infants, and infants learn to speak by mimicking what they hear. Eventually, this mimickry transmogrifies itself into genuine skill; rather than simply repeating phrases or words that have been heard, the infant -- child -- generates new sentences on his or her own.

This is a great object lesson in how unreliable common beliefs really are, because it's a complete load of crap. The first and most obvious problem with this story is that the process by which mimickry -- which parrots and other birds can do -- becomes actual skill in using language is not described at all. One may as well say "and then a miracle occurs!", because on this story, there is no explanation as to how ordinary skill with language becomes present. The second problem is more subtle, but also more serious. This is the fact that infants are simply not exposed to enough language quickly enough to account for their developing facility with using a language. In other words, kids learn to talk too quickly to be just repeating what they've heard other people say. This has lots of names; the one I like is "poverty of stimulus". This means that the supposed stimulus -- the thing that's supposed to get infants to be able to talk, namely people talking to them -- isn't rich or full or detailed enough -- hence, "poverty" -- to produce the ability to speak.

It's not entirely clear what the right story is (although it is very clear that the mimickry plus miracle equals skill story is flat wrong), but the best one I know goes like this. We are all born with a set of possible languages we can speak. Think of it like a set of switches in the brain. The brain comes with the switches, that is. When we are exposed to language from other speakers -- parents, grandparents, siblings, random strangers, the television -- some of those switches are flipped off. That is, the ones which would allow us to produce and hear sounds in our language doesn't us, the ones which structure sentences in ways not needed for our language, and so on. So, we are born with the ability to learn a language, and stimulus from others enables that ability, turning it into a real skill, by teaching us which of the things we are born able to do we actually need to be able to do.

Learning all things could be just like that. We could be just activating a set of innate capacities. Which would mean that learning actually does not extend or improve our cognitive abilities.

What I need, then, is a clear case of improvement in cognition, one which can't -- or can't readily or plausibly, at least -- be explained as the activation of something innate. And the best way to get that case on the table is bring back a topic that I let drop earlier, and talk some more about reasoning.

The Forever Man - Part 1 - To Serve is to Live - Page 2

As Petros pulled himself over the top of the wall, he saw the rest -- his squad, a few others -- on the opposite side of the gulch. Far below lay a few dozen freshly-killed wasps, their wings still twitching, spasming their last on the corpses of earlier brethren. Nothing had yet made the top of the outer wall, and he could see no other new kills between the walls. A quiet raid, then.

He hailed the other side, and, receiving no response, hailed them again. A Pulmek turned and jogged out along the leftmost cable, easily balancing, skinsuit barely flaring to course-correct. He recognized her quickly as she drew near -- Swith, one of his Agnants. If he were honest, one of his favourites, if for no other reason than her ability to think for herself. It was a rare gift, he knew, among the lowest of the Pulmeks. There were reasons they had stayed as Agnants, after all.

Swith stepped off the cable lightly, and inclined her head. "Lefent."

"Swith," Petros said, returning the nod. "How go things?"

"They go as they go." She shrugged. "A handful of beetles. You’ve seen the wasps. We’re not sure what they’re looking for."

He frowned. "They seek food, as always."

"No. With respect. But no." She pointed over the top of the outer wall. "We’ve seen a herd on the horizon. Cows, we think, but can’t tell. There’s too many -- the sand covers them."

"Then that is why the bugs came. They aren’t hungry. They’re running away."

Swith considered for a moment. "That could be. Do you need to look for yourself?"

Petros nodded, restraining a sigh. "By Jary’s order. I’ll see this herd first. Have them," he gestured over the inner wall, "bring me a glass. I need to know what they are, exactly, and where they are going. We could use the food."

Without waiting, knowing as he did that Swith could be trusted, Petros crossed the cable to the outer wall, his skinsuit flared more frequently than hers, helping him stay balanced. He nodded to the Pulmeks there, noting those who were his -- Wiht, Alp, Kuwat, old Trum -- and overlooking the ones that were not.

He looked out over the wall, ignoring the few bugs who seethed at its base, skittering and murmuring to themselves; ignoring as well the occasional discharge of the guns as they sought prey. Beyond the wall lay the desert, the trees of the enclave quickly turning to shrubs, then grasses, then thin and scant reeds. And the desert stretched across the horizon, as far as any eye could see. He knew what was out there, was one of the few living Pulmeks who had been deep into the desert’s heart, rather than scavenging on the fringes. And he hated it, with every fibre of himself.

Thursday, June 21, 2012

Pressed for time.

So, no update today. However, there will be an update tomorrow (unusual, as I generally only update Monday to Thursday).

Wednesday, June 20, 2012

Practical Philosophy - Chapter 1 - Logic and Language - Page 3

It's a reasonable thing to worry about. Philosophers have, rightly, been accused of developing theories without concern for their applicability. (Of course, it doesn't work against all philosophers; many, particularly those who work in political philosophy, the philosophy of mind, and the philosophy of physics are very informed by the relevant practical implications of their work.) I wouldn't say that pure theoretical work is a bad thing, but if the claim is that logic helps us to think better, then it would be pretty damning if we actually can't think any better than we actually do.

Now, there are clearly good reasons to be suspicious of any general claim that human cognitive ability cannot be significantly altered by human activity. It happens all the time, after all. It's called "learning". Whenever you learn how to do something, you alter the way that you cognize. This can be profound, as when you learn how to read, and thus become unable to not see the marks on this page as meaningful. It can also be minor, as when you learn who Winston Churchill is, and thus become unable to hear the name without a handful of associated concepts being brought to mind. (For me, it's cigars, waistcoats, and bulldogs.) So, if this is what the criticism amounts to, then it's patently silly.

But, I think there is a better way to take the point, namely that any theory of logic which aims to improve our cognition is going to run into limits which have been established by the evolutionary processes which produced our human brains. There are some things we're just not going to be able to do. Probability calculations, for example. We're very bad at estimating probability; in fact, we're systematically bad at it, bad in ways that can be predicted in advance. Casinos make their money by exploiting those systematic biases. Similarly, we are bad at acknowledging our own faults, and bad at acknowledging others' strengths. If something good happens to us, it's because we deserve it; but if something bad happens to us, it's simply luck. Contrarily, if something good happens to someone else, that's just luck; but if something bad happens, well, then they had it coming.

We can build tools to accommodate for these biases, once we see that they are there. We can use calculating devices, like computers, to figure out probabilities for us, and learn to rely on their results. We can also work to counteract our tendencies to laud ourselves and to denigrate others through a variety of means; for example, we can insist on consensus between a number of people, thus reducing the chance that the resulting conclusion has been unduly swayed one way or the other. [Note: I'm ignoring the issue of trade-offs because it just complicates the issue, but they are a concern, too. When we rely on tools to counteract systematic cognitive biases, we do so at a cost, usually a cost of speed. This is particularly true when we end up relying on the work or thoughts of other people; as anyone who's tried to forge a consensus knows, the more people get involved, the longer it takes to reach a decision.]

Notice, though, that these are all ways of counteracting systematic problems in human thinking. They aren't ways of improving it; they're workarounds, alternate routes that try to avoid the ways in which our thinking leads us away from truth and into error. So, logic, as I'm defining it, is more. It's not just a workaround. It's not a counteracting influence. It's an actual improvement in our ability to think. And that, I think, is the real concern that could be raised about my proposal: that I'm suggesting we do something - implement a significant improvement to our ways of thinking - which we actually cannot do.

The Forever Man - Part 1 - To Serve is to Live - Page 1

"Are you listening?!"

"Yes, I am listening!"

"Then do as I say!"

Lefent Petros nodded and turned towards the wall. The habit of duty was an old one, as old as he could recall, and consequently he was loath to break it. But somehow Genal Jary always found a way to encourage doing so. It was partly his attitude, that snapping, brittle irritation, unknowingly perfected over years of failed challenges and lost opportunity. But, mostly, it was the fact that he had no idea what he was doing; had risen far beyond his ability, and stayed there through connections and luck rather than any genuine merit.

Today was a perfect example. Petros and his group had been the first at the wall when the attack came, sirens blaring throughout the enclave, shattering the forest’s tranquil peace. The wall had never been breached to his memory, had only been scaled twice, and each time the bugs had been driven back to their deserts to wither and die. And both times, he, Lefent Petros, had led the charge. Both times, he had stood shoulder to shoulder and back to back with Commer Zdeti himself, hammering their way through the writhing hordes of the worms, the skittering legions of the beetles, skinsuits flaring with power, songs of battle in their hearts. Fighting to protect both the few remaining living and the ranks of the helpless dead.

Genal Jary had, as far as he remembered, stayed on the wall itself, firing the guns. An important task to be sure; but a leader’s role was to lead, to fight, to inspire, and not watch and kill from safety.

Still, Jary was Genal and could not be denied. He had ordered Petros to scale the wall -- an easy enough task -- and report back what he saw on the other side. It was beneath him, a task worthy of an Agnant at best. He had wanted to point this out, in some more diplomatic way than he otherwise might. But Jary had decided to insist. And so, Petros would climb the wall, would report back what he saw, would do exactly as he had been told.

But no more.

The wall itself was vast, filling the entrance to the valley completely, and made of the same warm, faintly electric metal as the few remaining machines in the enclave. It was, to be accurate, a double wall, with a gulch in between, with narrow cables strung taut between the tops. The cables would be difficult for a bug, even the flying ones, to land on safely, never mind cross, but a Pulmek warrior at full charge could easily keep his or her balance even while running. And Petros knew, from experience, that it was technically possible -- although somewhat risky -- to leap from one to another. At the four corners of the wall stood guns, powerful tools that kept the flying bugs at bay. At least, they were supposed to. More and more frequently, they had to be manned, as the auto systems could no longer target accurately. And Pulmeks were fast to react, but nowhere near as fast as the machines.

Tuesday, June 19, 2012

Practical Philosophy - Chapter 1 - Logic and Language - Page 2

Take my claim about the purpose of logic. It might be argued, rather crudely, that we shouldn't try to alter the natural habits of thought that have been bequeathed to us, either by evolution or by divine hand. The "divine hand" angle is actually the more plausible one. After all, if a perfect being gave us certain ways of thinking, it's at least reasonable to suppose that this being did so better than we are able to. Of course, the whole "divine being" hypothesis is unbelievably implausible, but more on that later.

The evolutionary angle is rather silly, really. Evolution is, at best, a tinkering process. Think of it this way. Imagine that you have a boat -- it doesn't really matter what kind of boat, but let's say it's a big ol' Viking warship. That's your boat. And this boat serves you well for a while, but then problems crop up. It's difficult to find a dock for your boat. The oars aren't all long enough, so you don't move as quickly when becalmed. There's a rip in one of the sails. And so on. Now, there's two ways you can try to improve your boat. You can replace it -- get rid of the old familiar boat, and get a brand-new one, without all the old boat's problems. Or, you can alter the parts that aren't working well for you.

The latter is pretty much how evolution works. Since it doesn't have the option of getting rid of a species and replacing it with a brand-spanking new one, evolution always takes what it has -- the resources available, biologically and ecologically -- and makes minor changes to them. That's how evolution is a tinkering process: it takes what there is, and makes minor changes to it. However, the boat case is only "pretty much" like evolution. When we're talking about your Viking warship, you're looking for the best Viking warship possible. When we're talking about evolution, we're not talking about the best species possible. The goal of evolution -- insofar as it makes any sense to speak of a biological process having a goal -- is to produce creatures that are better adapted to their environments. And "better-adapted" just means "less likely to get killed or drop dead before they reproduce and raise offspring to viability". (That last clause is sometimes optional; some offspring are not raised to viability by their parents, but have to figure it out on their own.)

So, it's silly to think that the way our brains work -- if we take that as a result of evolution -- as being the best way for us to think. There are going to be limits on how we think, of course; we're not going to be able to think in ways our brains can't physically handle. But it's just wrong to say that evolution has given us the best possible means of thinking imaginable. Evolution never reaches for the best possible anything -- it settles for what's good enough to get a species to successfully reproduce. And evolution always tinkers -- it may improve our brains, but it will only do so marginally and fractionally; big honking leaps forward in our ability to understand and know will have to come from somewhere else.

That "somewhere else" would be us, incidentally.

A more sophisticated counter-argument is available, though, which would hold that it is actually not possible for our brains to work in a way that is consistent with logic. That is, accepting that evolution has given us brains that work for "good enough" purposes, rather than perfect ones, we're never going to be able to think in a way that's much better than the way we actually do. So, it's not that we shouldn't mess with evolution's results; more that we can't. And any view, such as my own, which is contrary to this is dangerous, as it holds up a standard humans just can't live up to.

The Forever Man - Prologue - Page 4

Mary stood up. She was pale -- which was fair enough, Rayner supposed -- and swayed as she got to her feet, but otherwise had not a mark on her. Never mind severe injuries, the sort one would expect from a fall of hundreds of metres; she wasn't visibly bruised or scratched at all. Her clothes were torn and shredded, likely from the pane of glass that fell with her, and Rayner could see skin clearly underneath. Mary was entirely unhurt. There wasn't even any blood.

In fact -- Rayner peered more closely at the image -- she was not only unhurt, she was in fact unmarked. Rayner knew a fair bit about the medical history of her employees; had needed to, in order for the experiment to be at all possible. Mary should have been marked in a few places. She had fallen about two years ago, while taking a short skiing vacation, and cut herself badly. The clinic where she was -- Switzerland or something; Rayner couldn't remember -- had been poorly-equipped, and, while fixing the worst of her hurts, had left her with a ten-centimetre scar over her right shoulder. She also had a very faint, but very long, scar down the outer side of her right calf, stretching from the knee almost to the ankle. The result of some childhood trauma or another. But what Rayner was noticing -- the reason the Machine was filling her mind with the location of every slight imperfection on Mary's body -- was that none of them were there. No scars. No tattoos. No freckles. No ingrown hairs.

She seemed, in short, to be physically perfect. Or, perhaps more accurately, perfected.

Rayner turned away in her chair, the image flickering off as she did and the surface of the Machine going back to black. The best she had hoped for was that Mary would survive the fall, possibly with crippling injuries. Then, she could have returned to R&D, improved the process, tried again -- and again, and again, as many times as necessary -- until she had reasonable assurance that even serious bodily insults could be survived. That had been the goal, after all. Survival after otherwise fatal harm, for those who could pay.

But never had she believed that something like this would happen. Healing all injuries. And doing so perfectly, to the extent that even old injuries, old wounds, blemishes and imperfections had been removed. She had to be sure, though. She flicked a thought at the Machine, to send a group of technicians down to retrieve Mary for further testing. And, with another thought, added a group of guards. Not that Mary was likely to resist, but you never quite knew what you'd run into outside the tower on a night like tonight. The data the technicians would collect was too valuable to risk.

In fact, Rayner decided, the process itself was too valuable to risk. She couldn't rush this. Not that she hadn't done that before, pushing a product out to the market before it was ready, consequences be damned, but this was too important. It had to be confirmed and verified. Every step in the process had to work, every time, with every possible biochemistry. There could be no flaws, no accidents, no long-term side-effects. It had to work.

Because, if it did -- if she had actually achieved what she thought she had achieved…

"This changes everything," Rayner said to the empty room.

Empty except for the Machine.

Monday, June 18, 2012

Practical Philosophy - Chapter 1 - Logic and Language - Page 1

Logic is a really broad topic. We can talk about logic generally, as a synonym for "reasoning". We can talk about logic in the sense of mathematical logic, where what we're looking at is a system of rules for manipulating symbols. (This is usually the sense of logic in play when you talk to academic philosophers about it.) We can talk about logic as an attitude, something cold and emotionless, the contrast to passion. Language is also broad. If you go and ask a linguist, he or she'll tell you about grammar and phonetics, both of which are components of how language is structured, how we put sentences together, whether spoken or written. If you ask a philosopher, you'll get a discussion of semantics and pragmatics, both of which relate to what language does for us: respectively, what it means and how it's used.

And that's all fine and good. These are useful topics to think about and explore, and I'll probably get into some of both before the chapter is done. But the point of this chapter is to consider what logic and language are for. That is, what purpose, or purposes, they serve; whether better purposes are available; and whether there are ways to improve our logic and our language to achieve important goals.

So, let's start with the purposes. I'm going to give two ideas that seem, to me, to be pretty plausible as the purposes of logic and language, respectively. Both are somewhat controversial and will require refining. Once that's done, I can then move along and talk about the ways to serve those purposes.

Here's the first stab at what the purposes are. Logic is supposed to help us think better. The reason for having logic, and the importance it has for us, is to take the grabbag of psychological shortcuts evolution has given us, all the weird little cognitive habits human beings have developed over thousands of years, and refine them. Cut out or block the stuff that leads us away from knowing and learning, and improve the stuff that gets us closer to understanding our world. This means that logic's purpose is basically solitary. Me improving my logic doesn't help you out, nor vice versa.

Language, on the other hand, is essentially social. There's no such thing as a language without someone to talk to, someone you can use that language with. That probably seems a little obvious. It's equivalent to saying that the purpose of language is to communicate. Since communication requires at least one other person to communicate with, it follows that language requires some sort of social interaction. However, that's not quite the idea that I hold. I think that language is required in order for social interaction to exist at all. So, the purpose of language is to make it possible for us to be more than solitary creatures.

Those are the two claims I want to defend first. Logic is meant to improve an individual's ability to think. Language is meant to make it possible for individuals to form societies. There are, however, lots and lots of reasons to think these ideas are not only wrong, but dangerous.

The Forever Man - Prologue - Page 3

Now, all she had to do was wait. Either it had worked, or it hadn't. Either all her planning and preparation, her careful selection and cultivation of Mary, had led to the greatest success humanity would ever know. Or it hadn't, and she'd have to try again. Fortunately, there were lots of Marys in the world.

She didn't have to wait long. No more than ten minutes had passed when the Machine drew her attention to its surface. It did so forcibly, in fact, firing synapses and neurons without her intervention, causing her to lean forward and look down. Rayner did not appreciate the gesture. She knew it hadn't done that before -- in fact, she hadn't known it could do that. She was used to the Machine's soft voice in her head, its gentle guidance through her thoughts; to be moved around physically by its command was new, and deeply unpleasant. She resolved to have a quiet word with R&D tomorrow, to have that capacity removed. The Machine was impressive, but it was still just a machine. It would be better if it remained a tool. After all, when you swing a hammer, you don't want the hammer to hit back.

The Machine's surface served many purposes; for now, it served as a monitor, allowing her to look through the thousands of cameras and other recording devices that littered the tower and surrounding streets. The monitor focused on a knot of people about forty metres from the tower's front entrance, clustered tightly together, a small island of calm in the mayhem of the New Year. The scene skipped, and focused again, the Machine shifting cameras as it reached the limits of one's range. She could see more, see that the people in the heart of the knot were looking down at something, those on the periphery struggling to get inside and see. It didn't take much to guess what that was; but shouldn't there be blood? The force of the impact should have reduced Mary to little more than pulp. Unless...

The scene skipped and focused again. She could discern faces clearly now, blank expressions, neutral and fairly calm. No horror. No disgust. Not even interest. The lives of the worker drones that served her interests were visceral and animalistic, she knew, but surely even they would fear death.
The scene skipped once more. And Rayner knew that she had been wrong. The expressions she had thought were neutral, calm, disinterested -- they could look that way, to casual observation. But at this distance, with the power of this camera, looking at the focused images carefully chosen by the Machine, she could now see the fixity of their gaze, the set of their jaws, the pale skin, the sweat and tears running down.

And then they broke -- the knot disintegrated in a matter of seconds, the people scattering, fleeing, shoving their way through the crowds. She saw a few shove back, without effect. The only concern was escape from -- and Rayner looked back at what they had been surrounding, what they had seen.