≡ Menu

Musings on Impartiality

Marc Millis, Tau Zero’s founding architect, drawing on his experience with NASA’s Breakthrough Propulsion Physics project and the years of research since, offers us some ideas about impartiality and how scientists can hope to attain it. It’s human nature to want our particular theories to succeed, but when they collide with reality, the lessons learned can open up interesting alternatives, as Marc explains in relation to interstellar worldships and the possibilities of exotic propulsion.

by Marc G. Millis

The best researchers I know seem to be able to maintain their impartiality when reaching new conclusions. The more common behavior is that people get an idea stuck in their head and then try and prove themselves correct. I just learned that there is a term for this more common behavior: “Polemical.” Embedded in the word is the notion that controversial argument can turn aggressive, an inevitable result when people are defending what they consider their turf.

I mention this in the context of getting trustworthy results, and then acting on those reliable findings rather than just charging ahead based on unverified preconceived notions. If the overall intent is to make the best decisions for the future – then decisions rooted in reliable findings, rather than expectations, will be more in tune with reality. They will be better decisions.

The topic of interstellar flight affords opportunities for easier objectivity as well as the opposite – pitfalls where one can lose objectivity. Because interstellar flight is almost certainly farther in the future than the next Moon and Mars missions, it is easier to apply impartiality. The huge payoffs of interstellar flight (finding new human homesteads and new life) are far enough away that there is no need to sell a particular pet technology today or skew the results toward near-term promises.

That said, I surprised myself when my own assessments gave results different than my expectations. Case in point – estimating how far in the future the first interstellar missions are, based on energy (energy is the most fundamental currency of motion). Those findings and a refined sequel (to appear soon in the Journal of the British Interplanetary Society) indicate that the first interstellar missions might be 2-centuries away, albeit with huge uncertainty bands.[ref]Millis, Marc. (2010) First Interstellar Missions, Considering Energy and Incessant Obsolescence, JBIS Vol 63 (accepted, pending publication).[/ref] The first Centauri Dreams post on those findings met with ‘energetic’ reaction, where many seemed disappointed that the prospects seemed so far in the future. Before I ran the numbers, I suspected that it would be much sooner too. The first calculations were done around 1996, and those results made me rethink what ‘next-steps’ were really required.

Rather than proceed with my prior notion, I had to stop and rethink things. The data said something unexpected. I knew I had conceived the methods to be impartial and fed the assessments with unbiased data, so the findings would be similarly unbiased. They were what they were. So, should I redo the analysis until I got the answer I wanted, or accept the results for what they were and then re-adjust my expectations? I decided to expose those results to other reviews, to check for errors and such, and then to accept the findings as they were.

Before that point, I thought the next step would be to use more detailed energy assessments to help pick the best interstellar propulsion options, but with two centuries of time to plan ahead, and many options whose numbers were still debatable, I realized that we need to abandon the idea of trying to pick the ONE best interstellar solution. Instead we need to focus on getting reliable data on the wide span of ideas (no salesmanship) – and to investigate the most critical ‘next-steps’ on as many of them as possible.

And this long lead time provides the topic of interstellar flight with the opportunity for more objectivity – the opportunity to take our time to reach sound decisions – to provide more trustworthy progress.

Colony Ships and Spaceship Earth

The other result that I was not expecting was that colony ships might be easier to launch than small, fast probes – at least in terms of energy. My prior expectations were that colony ships would need to be so immense and complicated that they would take longer to develop than a fast probe to Alpha Centauri. The energy study showed otherwise. Kinetic energy is linear in mass, and goes as the square of speed. That means if the ship is twice as massive, it requires twice as much energy, but if it goes twice as fast, it requires FOUR times the energy. Colony ships do not need to go fast. They only need to drift, carrying a segment of humanity. Up to that point, I thought colony ships would be a sequel, not a prequel to small, ultra-fast probes. Sometimes you just have to run the numbers.

Then it occurred to me, while I was drafting my first TEDx talk, that the notion of such slower interstellar world ships also provides a more impartial venue to discuss critical human survival questions. Colony ships allow us to consider these questions with NO dependence on conventions or biases. If designing a society from scratch, one is free to start anew to fit the facts as they are discovered. On Earth, however, when dealing with questions of population size, environmental stability, amount of territory per person, and governance model, the debates are typically won by cultural edict (e.g. no birth control) or warfare (quest for territory or power). So, after all that, I realized that colony ships merited far more attention than I originally gave them, and hence, we will need to track down suitable pioneers to cover those issues too as part of Tau Zero.

Unbiased Physics

When it comes to one of my pet topics – propulsion physics and the quest for space drives – I ran into another facet of impartiality. I found that many physicists do not like to work on problems with potential applications since the application ‘taints the purity’ of the research. Instead they want to be driven by curiosity alone. In other words, they do not want to be biased. In the quest for propulsion physics, where I really hope a space drive method can be found, I have an ingoing bias. I want the results to turn out a certain way. This creates a conflict of interest in how I might view – or skew – the results. To make genuine discoveries, however, I must discipline myself to avoid imposing such biases. Although I can let my wishful musings help me pose the key questions, to get real progress I must also let the findings – unbiased findings – answer those questions. I must accept the results as they unfold.

Take the case of black holes in contrast to traversable wormholes or even warp drives. Studying black holes has revealed insights about spacetime warping, presumably without bias since no desired result is sought. But if one studies the very same physics in the context of faster-than-light wormholes or warp drives, one might get biased results because of wanting such devices to be feasible. Fortunately, much research published on these topics has maintained the rigor to avoid the taint of such biases. Insights into spacetime physics are also being learned by pondering warp drives and wormholes. These questions are even presented as homework problems in textbooks (e.g. Hartle (2003) Gravity: An Introduction to Einstein’s General Relativity).

The irony is that even the curiosity-driven research has implicit biases – that natural sense of ownership that a person has for their research ideas. There is an urge – even in this case – to have the findings prove the author’s pre-conceived point. This is just a human norm. High-quality physicists can discipline themselves to separate out this bias. In contrast, I’ve also seen physicists discuss their ideas with the same possessiveness as kids with toys on a playground. Regardless of our motivation in searching for new knowledge, we must maintain vigilance to avoid imposing our own biases on the findings, even the implicit bias of self.


Comments on this entry are closed.

  • Astronist July 1, 2011, 18:54

    Kenneth Harmon,

    The security you are looking for does not require an extrasolar “habitable planet”. The destinations you are seeking are here in our own Solar System: firstly the Moon and Mars, then the asteroid belt. There is room for thousands of years of dynamic growth and diversification here before there will be any pressing necessity for our descendants to travel further afield.

    There are of course in any case a number of habitable planets in our own Solar System (habitable for microbes, and also habitable for technological species).


  • Eniac July 2, 2011, 12:50


    Re. The Toaster Project – that’s not quite accurate, since one would not be working alone. Also, attempting to replicate the exact design of the toaster makes it harder than it has to be… and a clay toaster, rather than a plastic one, is just that much cooler anyway. Besides, if you want to make an apple pie from scratch, you must first create the universe.

    Sure, these are good points. On the other hand, though, a toaster is not the most complex or important of artifacts. Colonists would not want to be without cell-phones and computers (in fact, they could not be), and those are far harder to make from scratch than toasters, even with an army of experts. A clay computer won’t cut it, no matter how cool it may be.

    Luckily, the universe has already been created, so if we allow “scratch” to mean regolith and rubble as found on most solid bodies in a typical star system, plus starlight for energy, that will suffice.

  • Kelvin Long July 2, 2011, 15:19

    Dear All,
    A very nice article from Marc Millis and it generated a lot of interesting discussion. Here is my own, slightly different, take on this issue.

    Marc said: “we must maintain vigilance to avoid imposing our own biases on the findings, even the implicit bias of self”.

    This is true but I also think there is a place for bringing obout something by having an optimistic vision that it can occur – this is what Clarke called ‘creating a self-fullfilling prophesy’. Otherwise you run the numbers, it says x years away, so you give in. Instead I think you have to take all projections with some pinch of salt when they go past >30 years or more. By trying you discover things along the way, create new technologies which act as a catalyst to technological innovation and makes the ambition that more likely.

    Also, kenneth referred to Marc’s projections of ~200 years as a ‘best case’ scenario. I dissagree – I think it is a ‘worst case’ scenario. When you look at all these projections done by many authors, looking at velocity trends, energy trends, economics…..they all tend to rely on linear scaling. I don’t think that is how the history of technological progress works – jumpting from one paradign shifting S curve to another for example. Three critical factors will change the linear scaling arguments:

    1) disruptive technology
    2) the arrival of AI, which is really a form of (1).
    3) compelling reasons for trying, i.e. exoplanets, immenent threats.

    The best thing we can do is a best case and a worse case projection. To my mind this is represented by a high-growth exponention technology increase and a low-growth or linear one respectively. When you run the numbers (which I have) it suggests that a medium-growth (>3% annual) or high-growth (>8% annual) exponential technology capacity increase will be required in order to facilitate the first unmanned interstellar probe launch by the year 2200 or 2100 respectively. This is assuing that to get to the nearest stars in ~100 years you need to be hitting a cruise velocity of ~2700 AU/year post-acceleration phase. This is the subject of a paper I have just submitted to JBIS in response to the paper from Marc Millis and others published previously.

    Like many others I cannot demonstrate rigorously when the first unmanned launch will occur. We cannot predict the future, all we can do is perform scaling calculations from current or near-future technology. I agree that ~2200 looks credible on paper for the date of the first launch. However, I also ‘believe’ this is pessimistic and that a first launch date nearer ~2100 may in fact be possible. Remember, what constitutes an interstellar probe? It may only be ~few kg science payload, that would still count.

    So, I admit to being biassed in my ‘belief’ that it is possible sooner than ~2200. But this belief strengthens my own resolve to make it happen and therefore increases the chances that my contribution can help it happen. It is quite possible that our work input to making the first launch happen is directly correlated with how far away we thing it is likely to occur. If the numbers all said it wouldn’t occur until 1,000 years from now, possibly some of us would change ‘hobbies’. Its believing it may be closer than some claim that inspires me personally to reach for that ambitious goal and see it happen, perhaps in my childrends life-time. I find this much more of an exciting reason to participate in intertsellar studies.

    Finally, Bob Forward wrote a paper in 1976 in JBIS titled ‘A Programme for Interstellar Exploration’, vol. 29, pp.611-632. Part of his abstract reads:

    “Annual funding for this pase of the programme wold climb to the multi-billion dollar level to peak around 2,000 AD with the launch of a number of automated interstellar probes to carry out an initial exploration of the nearest stellar systems. …..Assuming positive resurns from the probes, a manned exploration starship would be launched by 2025 AD, arriving at Alpha Centauri 10-20 years later”.

    Something for people to ponder, given that Forward knew what he was talking about.

  • Astronist July 2, 2011, 19:00

    “An optimistic vision that it can occur” — think of Brunel and the Great Eastern, or Von Zeppelin and the rigid airship, to give two non-space related examples.

    I agree, Kelvin, but would go further: there’s a strong political — or polemical — element in this, in that we are saying that it’s not only possible but desirable for us to get out there and become a spacefaring species.


  • Rob Henry July 2, 2011, 22:34

    Tobias Holbrook, are there really no comprehensive studies already on the necessary size of the seed for modern industry. This surprises me unless you just mean none with special reference to interstellar travel. This situation sounds like we have a death wish. What if we were hit tomorrow by a giant solar flare or comet or plague! Such a study would also be of great interest even its implications were confined to purely academic outlooks.

  • Tobias Holbrook July 3, 2011, 12:20

    “Tobias Holbrook, are there really no comprehensive studies already on the necessary size of the seed for modern industry.”
    None that I’m aware of. Any idea where they would be, if the studies have already been done?

  • Eniac July 6, 2011, 0:37

    Rob, Tobias:
    The best reference I know is Freitas’ book “Kinematic Self-Replicating Machines”, in which the subject of the size of an industrial seed is addressed and which contains loads of references to more detailed material. It is available on-line: http://www.molecularassembler.com/KSRM.htm.

    Also, Freitas’ work on interstellar probes is worth mentioning: http://www.rfreitas.com/Astro/ReproJBISJuly1980.htm

    Still, there appears to be a remarkable lack of attention to this subject, which is to be deplored.

  • Tobias Holbrook July 6, 2011, 16:23

    Hmmm, that was written in 1980; since then, we’ve become able to shrink the needed mass considerably I believe (for example, even simple 2d printers capable of manufacturing solar cells and intergrated circuits slash the mass required for those items to be manufactured). Say, it was approx. a 1000 tonnes in the second link, for a fully automated system… I’m thinking, perhaps 500 tonnes could do the trick with the advances we’ve made?

  • Eniac July 6, 2011, 23:07


    I fully agree with you an the impact of flexible manufacturing technologies such as printing. Even less high-tech methods such as casting and machining can be turned into highly versatile production facilities with the help of CNC, more like an automated workshop than a factory. The more versatile the production facilities, the less are needed to manufacture the full complement of parts. The more versatile the parts, the less are needed, period. These are the two variables that would most drive the design of replicators, and the industrial seed.

    The seed would be a “spore”, with a minimal subset of parts and a minimal subset of production facilities, just enough to bootstrap into an arbitrarily large “adult” version on site, presumably in multiple, carefully choreographed stages. Not unlike the bootstrapping of computer operating systems, only in the real world and much more complicated…