Выбрать главу

Let him feed the ducks.

“Floyd?”

“What?”

“You were miles away.”

“Years away,” he said. “Not quite the same thing.”

She steered him towards an ice-cream stand. Floyd dug into his pocket for a few coins.

TEN

Auger awoke to the rapid metallic popping of thruster jets, like a rivet gun. Her first thought was that something must have gone wrong, but Aveling and Skellsgard both looked alert and focused rather than alarmed, as if this was something they had encountered before.

“What’s happening?” she asked groggily.

“Go back to sleep,” Aveling said.

“I want to know.”

“We’re just dealing with some tunnel irregularity,” Skellsgard said, using her free hand to point to the contoured display in front of her joystick panel. She was flying now, while Aveling took a rest. The moving lines on the display panel were bunched and crimped together. “The walls are pretty smooth most of the way through, but every now and then we come across some structure or other, which we have to steer around.”

“Structures? Inside a wormhole?”

“It isn’t a wormhole,” Skellsgard began. “It’s a—”

“I know: it’s a quasi-pseudo-para-whatnot. What I mean is, how can there be any kind of structures inside this thing, whatever it is? Isn’t it smooth space-time all the way through?”

“That’s what you’d expect.”

“You’re the theorist. You tell me.”

“Actually, there’s a good measure of guesswork involved here. The Slashers didn’t tell us everything, and they probably don’t have all the answers themselves.”

“So give me your best guess.”

“OK. Theory one. You see these stress-energy readings? They relate to changes in the local tunnel geometry ahead of us.”

“What are you sensing them with? Radar?”

Skellsgard shook her head. “No. Radar—or any EM-based sensor, for that matter—doesn’t work too well in the hyperweb. Photons are absorbed into the walls or scattered chaotically by interaction with the pathological matter. And looking ahead is like trying to see sunspots with your naked eye. Neutrinos or gravity-wave sensors might work better, but there isn’t enough room for them in the transport. All that’s left is sonar.”

“Sound?” Auger asked. “But we’re moving through a near-perfect vacuum, aren’t we?”

“As near as dammit, yes. But we can persuade a kind of acoustic signal to propagate through the lining of the walls. It’s like the compression wave that the transport’s surfing, only about a billion times faster. It propagates through a stiffer layer, a different phase of pathological matter with a much higher rigidity. It’s how we send signals down the pipe, so that we can talk to the portal at the E2 end. Trouble is, it doesn’t work when a ship is in the pipe: we act as a kind of mirror, bouncing any signals back the way they came. But we can send our own signals up the line. They’re not strong enough to reach all the way to the far portal, but they do act as a kind of feeler, sounding out obstructions and irregularities in the walls.”

“That still doesn’t tell me what causes those irregularities in the first place.”

“Here, take a look at this,” Skellsgard said, directing Auger’s attention to a knot of very close contour lines oozing into view on the display. “This is the computer’s best guess at the shape of an approaching irregularity in the tunnel lining, based on the echoes from the sonar. If the contours were bunched together symmetrically, we’d be looking at a constriction, a narrowing in the tunnel ahead of us. But that isn’t what’s happening here. There are places where the tunnel lining looks as if it’s been etched away, and places where it bulges inward. Theory one says that this is symptomatic of some kind of decay of the basic fabric of the link, either due to lack of maintenance or not enough ships using it.”

“Not enough ships?”

“It could be that the ships are meant to perform some repair function when they pass through. That’s what we call the ‘pipe-cleaner hypothesis.’ ”

“Fine. What about theory two?”

“This is where it starts to get seriously speculative,” Skellsgard warned. “Some people studying the link have made records of these irregularities, accumulating data from many transits. Of course, the data is very noisy and subject to the interpretive vagaries of the navigation system. So then they take those records and feed them into maximum-entropy software to squeeze out any latent structure. Then they take the output from that process and feed it into another bunch of programs designed to sniff out latent language. One such procedure is called the Zipf test: it involves plotting the logarithmic frequencies of the occurrence of different patterns seen in the walls. Random data has a Zipf slope of zero, whereas the Zipf slope of the tunnel patterning is pretty close to minus one. It means that the signals in those walls are significantly more meaningful than—say—squirrel-monkey calls, which only get down to minus point six on a Zipf plot.”

“Not conclusive, though,” Auger said.

“But the researchers don’t stop there. There’s another statistical property known as Shannon entropy, which even tells you how rich the communications are. Human languages—English, say, or Russian—have Shannon entropies around the eighth or ninth order. That means if I say eight or nine words in one of those languages, you can have a pretty good stab at guessing what the tenth is going to be. Dolphin calls have Shannon entropies in the range of three to four, whereas the tunnel scrawls are up at seven or eight.”

“Less complex than human language, in that case.”

“Granted,” Skellsgard said, “but their true complexity might be masked by the errors we introduce in decoding the sonar images. Or the messages themselves may be blurred by erosion or some other process we don’t understand.”

“So theory two is that the patterns are deliberate messages.”

“Yes. They might be analogous to old highway signs: speed limits, temporary restrictions, that kind of thing.”

“You’re not serious.”

“You haven’t heard anything yet, Auger. Want to hear theory three?”

“Oh, why not?”

“This is definitely not accepted wisdom, I should warn you. Theory three says that the tunnel patterns are a kind of advertising.” Auger opened her mouth to say something, but Skellsgard kept on talking. “No, wait. Hear me out. It makes a warped kind of sense when you think about it. Why wouldn’t a galactic supercivilisation have advertising? It seems to be pretty much glued to our culture, after all.”