Apologies if the title is confusing - I've been trying to figure out how to word it for a while now.

Essentially I am writing a (relatively) hard science-fiction story set in the present/very near future, where a strange phenomenon is causing stars in a distant region of the Milky Way galaxy (roughly 18 to 25,000 light-years from Earth) to accelerate 'upward' relative to the plane of the Milky Way's ecliptic. This 'strange phenomenon' is due to the machinations of an alien race unknown to humankind, who I call the 'Shapemakers'. For an unknown reason, the Shapemakers have deployed a network of slower-than-light Von Neumann probes that have been spreading from star to star in this region of the Milky Way and dismantling any matter they come across (primarily that of debris disks and planetary systems in orbit of each star). This matter is then used to construct immense Class A stellar engines (based on the principle of the Shkadov thruster). The Shapemakers seem unconcerned about anything other than the stars themselves: their goal is simply to move as many stars as possible for unknown reasons.

I've done some preliminary research and it seems like a Shkadov thruster would impart an extremely small amount of acceleration compared to a star's relative velocity as seen from Earth, on the order of 10^-12 m/s squared of acceleration. It would thus take millions of years to achieve any kind of significant artificial deviation. At what point would we be able to tell this 'engineered' acceleration apart from a star's natural radial velocity, if we even could? My other inclination is that we'd be able to see a Shkadov thruster by monitoring these far away stars, given their immense size and the sheer amount of reflected light output. I assume they'd be detected relatively easily (hence why none have been sighted in real life yet), but is there any reason why it'd be difficult for humanity at its current technological level to spot a Shkadov thruster? The less noticeable this is, the better - I'd like this to be a relatively new discovery that requires specialised, cutting-edge technology that only a few have access to, rather than something any old Joe renting telescope time can take a screenshot of.

The second part - and main plot of the story - revolves around an alien spacecraft entering our Solar System from one of these distant stars. The Von Neumann machines began to dismantle their system several millennia ago by human standards, and said aliens barely had time to construct a kind of 'ark' before their homeworld was destroyed. The ark shows up a little while after humans have figured out the stars are moving, and takes up an orbit around one of the outer gas giants to start mining them for materials. Humans naturally want to make first contact and send an extremely expensive mission out to this alien spacecraft. The thing is, would it make sense for the humans to already know about the star movements, so the aliens just tell them why it's happening? Or would it be impossible for Earthbound humans to figure this out, and the aliens would have to tell us the entire thing?

There are other considerations regarding the aliens, but they're not relevant right now. Ideally, I'd like a situation where these stellar engines are relatively accurate to physics, but extremely hard to find - something that would be easy to cover up or miss - but not so hard to detect that it's literally not possible to find them.

Would this be possible, or am I going to have to rethink my premise?

  • $\begingroup$ Reflection of light: Allow me to introduce myself $\endgroup$
    – Alastor
    Commented May 13 at 11:47
  • 1
    $\begingroup$ Pardon a perhaps dumb question, but how are you dealing with the issue that anything anyone on Earth could detect would have happened at least 18,000 years previously? So any society carrying it out would be long gone by the time its detected. $\endgroup$
    – T.E.D.
    Commented May 13 at 13:42

7 Answers 7


My other inclination is that we'd be able to see a Shkadov thruster by monitoring these far away stars, given their immense size and the sheer amount of reflected light output. I assume they'd be detected relatively easily, but is there any reason why it'd be difficult for humanity at its current technological level to spot a Shkadov thruster?

A Shkadov thruster produces two things: an acceleration on its parent star, which is detectable through Doppler shift, and light.

But the light has to be directed towards the Earth to be detected, and the Doppler shift only applies to the component of the velocity that is directed towards Earth (or in the opposite direction). If the distance between Earth and the star remains constant, there is no Doppler effect.

Shkadov stars accelerating away from, Earth would appear as stars with higher luminosity than normal, and probably a strange polarization of the apparent output because the Shkadov statite would (partially) polarize the incident light. If the star has any variability, there will also be a detectable temporal autocorrelation in the star's apparent output that will even allow, given enough data, to determine the position of the Shkadov thruster.

Shkadov stars accelerating towards Earth would show less luminosity than normal, because the statite shields them, and an uncommon Gegenschein effect caused by the light exhaust interacting with the interstellar medium. The light from the statite couldn't be seen, so no polarization effects and no autocorrelation. I feel that detecting this kind of stars at galactic distances would be next to impossible.

What about Shkadov stars whose velocity is not directed towards Earth. These might fit your requirements. In vacuo, they would not be detected at all (no Doppler shift, as said above, and no observable exhaust). On the other hand the energy output of the Shkadov thruster could ionize interstellar gas behind the thruster, and would do so at the speed of light (this is called a light echo or nova flash effect). So, given a particularly sensitive telescope array, you would observe a "focused" light echo effect (this is a non-directional one) as well as, maybe, a noticeable Tyndall effect on the same medium.

The rationale behind the discovery might be a space telescope search for lasing stars, an effect that requires large nebulae. So, a survey of several suitable nebulas would be performed using the largest, most sensitive space telescope available.

"Non-isotropic light echo effect observed in [NEBULA NAME]" might be the noncommittal title of a paper from someone observing the phenomenon, and probably realizing what it had to be, but not wanting to risk ridicule by coming out with an alien technology hypothesis.

Then, someone else might focus the space telescope along the light echo axis, measuring the Doppler shift of all candidate stars - and would observe a growing discrepancy in the shift of a group of them.

"Anomalous time-varying Doppler shift observed..." could be the title of another paper on those stars.

Spectral anomalies would be identified, compatible with part of the star light being reflected by a large structure. From there, the likely "natural" spectrum of the star could be guessed, and from that you have the probable mass of the star. Knowing the mass and the Doppler shift rate of variation (which gives the acceleration) you can estimate the thrust, and compare it to the entity of the observed light echo.

This way you can frame the observations like, "The interstellar gas in the vicinity of star XYZ behaves as if it was reacting to a polarized beam of light from a large-scale, low-density reflector intercepting about 5% of said star output with efficiency above 95%, located at a distance of ... UA from XYZ. The star also behaves as if it was subject to an asymmetrical gravitational pull from a statite of the above mass and position".

At this point someone would utter the dreaded words, "Shkadov thruster".

Or, if popular media got their hands on it first, seeing their usual restraint in titling, we might start with "Evidence of a Stellar-scale thruster observed around BFT 1234567".

  • 1
    $\begingroup$ "Normally they would not be detected at all.": that's assuming they harness the full output of the star without losses. They may find it more practical to just grab, say, 90% of the light, allowing the rest to escape through gaps in their megastructure. This would give you a star with the spectra of one with 10 times the luminosity, which would be...odd. $\endgroup$ Commented May 11 at 23:20
  • $\begingroup$ I think the reflection off interstellar gas would be detectable because it's not blackbody. And even if it somehow was it would be way too hot. $\endgroup$ Commented May 12 at 4:00
  • 1
    $\begingroup$ @ChristopherJamesHuff yes, but the point is, this is in the hypothesis that the axis of thrust is directed at a significant angle. Then we'd not be able to observe a significant Doppler shift, since the star's velocity would not be directed either towards or away from us, and we wouldn't be able to detect the light beam or its Gegenschein halo from the statite for the same reason. But if the interstellar medium is dense enough, then we can see the effect on that. $\endgroup$
    – LSerni
    Commented May 12 at 16:57

One of the biggest problems your astronomers have is establishing the velocity vector of the affected stars at all.

The velocity component of your moving stars that is directly towards or away from the Sun will produce doppler effects that shift the spectrum of the star in a detectable and suspicious way. Movement in any other direction is significantly more difficult, because you actually have to see the things moving.

For objects which are close enough, you can use parallax effects to work that out, but this requires some pretty good telescopes. The currently limit for Earth-and-its-immediate-surroundings-based astronomy is ~18000 lightyears, and anything beyond that distance has a parallax too small to detect. Fancier telescopes with longer baselines extend that distance, which may be required if you need stars as far away as 25000 ly to have their speed and direction determined.

Determining acceleration is probably impractical, unless there was some way to measure historical velocity, such as being able to observe shockwaves in large gas and dust clouds that a moving star had passed through a long time ago from which it might be possible to work out that it is travelling the same direction but is now much faster. Tricky, though.

At what point would we be able to tell this 'engineered' acceleration apart from a star's natural radial velocity, if we even could?

It is already possible to do identify stellar associations with a bit of astronomical detective work. A bunch of stars with unusual velocity vectors that all seem to be aligned but aren't otherwise obviously associated and have no other obvious source for the motion is going to raise some eyebrows.

As suggested above, detecting acceleration rather than suspicious velocity is impractical.

is there any reason why it'd be difficult for humanity at its current technological level to spot a Shkadov thruster?

It isn't quite as trivial as you might think.

Dyson spheres may be detectable by their peculiar infrared spectrum. The solar sail portion of a Shkadov thruster should behave similarly weirdly, albeit on a somewhat smaller scale, but its thermal output will be swamped by the simple fact that it is right next to an actual star. A thruster moving directly away or towards us would look very peculiar, but with transverse movement the sail might not be detectable. All the current tricks for detecting strange things in stellar systems rely on the fact that those things orbit (which a statite like a Shkadov thruster will not) or interfere with tranmission of light between the star and an observer (which a transversely-thrusting system will not).

I don't think you'll have any problem handwaving in advanced astronomy techniques for spotting thrusters, and constructing telescopes further away from Earth to provide longer baselines and better parallax measurement requires advanced in spaceflight and so on.

would it make sense for the humans to already know about the star movements, so the aliens just tell them why it's happening?

That seems fine. I don't think the aliens have to provide all the exposition here, but providing an explanation of a phenomenon that's been detected and puzzled over by humans does make your job easier. Its up to you whether you want that sort of style of revelation, though.

  • $\begingroup$ With regards to the difficulty of observing motion in directions other than towards or away, would it be possible to just look at the shape of the star's bow shock? I mean, a start moving "Up" from the galactic plane should have a different wake in the interstellar medium than one moving with, against, or "Down" from the galactic plane. $\endgroup$
    – Ryan_L
    Commented May 12 at 2:02
  • $\begingroup$ @Ryan_L it depends on a bunch of stuff, like how big and fast and hot the star is, and how dense the medium it is travelling through is. Theyre not so easy to spot for very distant smallish stars. $\endgroup$ Commented May 12 at 7:33

Parallax measurements are very difficult at great distances. But perhaps a large telescope could be sent out of the Solar System using some new form of propulsion (fusion drive perhaps) for just such a purpose as the extended base line would make the measurements easier.

It is entirely plausible that such advanced aliens would have no interest in discussing anything with humanity.


Any sufficiently large structure capable of harnessing solar energy by order of Yottawatt scale, would produce

  1. A highly detectable infrared signature. Theoretical physicist Freeman Dyson postulated alien civilizations (Type II+ on Kardashev) scale would produce enormous infrared focused energy.
  2. Observation of high energy photons coming from one side of the star/other.

It's rare for activity of any kind on the other side of the galaxy to be detected due to the intervening dust lanes which block out light. Certainly you are safe from amateur instruments detecting anything significant from that direction. Longer infrared wavelengths can have more success.


I don't think this would be possible. 18000 light-years is really far away, and at that distance, the actual movement of the stars at the acceleration of a shadkov thruster would make such a tiny difference as to be impossible to detect. That means it can't be identified based on movement. The other possible way we might detect it is noticing a sudden increase or decrease in the brightness of the star, but this would likely be chalked up to gass clouds or the such obscuring the light. We would probably never think that it's because of an alien megastructure. The only possible way I can think of that might lead to this is if the mirror is between us and the star, and then we might notice the star disappearing, but then it would also not be pointed to accelerate the star above the ecliptic. The only way for that to work is if the star is directly "below" us, but that is only a few hundred light years away, not 18000. So no, it would not be possible for us to detect this with out current day tech.


A note on timescales

Other answers have given nice details about detecting the star movement, so I'll focus on a different part of your question:

would it make sense for the humans to already know about the star movements, so the aliens just tell them why it's happening? Or would it be impossible for Earthbound humans to figure this out, and the aliens would have to tell us the entire thing?

Let's have a look at the timescale. The aliens arrive here in an "ark" and you would like:

The ark shows up a little while after humans have figured out the stars are moving

This would have to mean, since the stars in question are 18000-25000 Lightyears away from us, and light takes a year to cross a lightyear, that the stars exhibited noticeable movement 18000-25000 years ago.

Since the probes dismantled the aliens' planet to build the thruster, the ark would have to have launched at about that time too. Since the aliens didn't arrive earlier, they either have "only" light-speed travel capacity, or they took some detours before getting here. And 18-25 millenia is a long travel time, so non-stop travel might be less probable than "fly to a different star, warn their civilization, fuel up from their gas giants, resupply from their asteroid belt, then move on, repeat". Or maybe we're the first civilization they find, and they just did the refueling bit.

Either way, if a mission to make first contact is "extremely expensive" to us, you're probably not looking at far-future sci-fi with very futuristic technology - while the aliens have managed to build a space ship that was able to travel for 18-25 thousand years, keeping their civilization alive through many, many generations (unless they're immortal?) and not failing catastrophically.

Humanity is smarter or more creative

It will take some suspension of disbelief that we, who can barely make contact with a space ship that's in our solar system, could solve a problem that the species who built that ship and kept it operable through 18 to 25 thousand years could not. But sometimes one just does not think of an option - so maybe the aliens teach us their technology, and we find a fun new way to apply it. Doable.

Did we already know about it, or did the aliens tell us?

Well, if the timeline is as I described it, we might or might not have figured it out already (see the other answers) so it's your choice.

But if you don't want your aliens to have taken thousands upon thousands of years to get to us, meaning 18000 to 25000 years have NOT passed yet, then the light of their stars that we can currently detect here on earth shows them peacefully and undisturbed following their regular, unaltered path. So if the aliens got here using FTL travel, they would have to tell us - since we could not possibly see it yet.

Alternative explanation

An alternative that would let the aliens not spend more time than we have properly recorded history on their ship and let humans already have theories or knowledge about what's happening:

The aliens saw their neighbouring stars change course. They observed it, maybe used their FTL drives to go take a look. They saw what's happening, saw that they had no chance to stop it, and started building their ark. When the first probe(s) arrived at their own system, they boarded the ark and started it.

Humanity can't see that their star has changed course yet, but you did say that many stars are changing course - so we just started noticing the first two or three stars changing course, while the aliens know that their star is the thousand and third one and is currently getting its thruster built from the mass of their home planet.

  • $\begingroup$ Does it help anything if Earth has a long time to accumulate data and then someone discovers weirdness in old records? Not 1000 generations, but a century or three. That's time to develop future tech, and to let it travel a bit (if that's needed). "Hm, that's strange, no idea what it means, let's build [something whizzy] to find out more...." This might expand the scene-setting introductory material to a sizable chunk of the story, which could be fun but maybe not the fun that the OP has in mind. $\endgroup$
    – Mark Wood
    Commented May 13 at 21:02
  • $\begingroup$ @MarkWood more time to accumulate data always helps (well, unless the time is interrupted by some "library of alexandria" catastrophe). We are still gaining valuable insights from old astronomical plates (basically photographs of the night sky, on glass) from 1880 onwards - see astronomy.com/science/… $\endgroup$
    – Syndic
    Commented May 14 at 5:54

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .