An important concept in earthbound geopolitics is the "natural boundary" -- which is to say a defensible frontier beyond which further advance becomes difficult given any sort of serious opposition. Natural boundaries are important because they are basic units of empire -- an empire expands past a natural boundary with difficulty, but once it has conquered a whole region within boundaries of this sort, can easily hold them. Natural boundaries thus tend to also become political boundaries, whether imperial or provincial, and where natural and political boundaries are identical, very stable boundaries.
If we postulate civilizations expanding past single worlds to become interlunar, interplanetary or interstellar in scope, then we should attempt to discover what sort of natural boundaries might affect the growth and defense of an spacefaring civilization. This article examines the possibilities, assuming "real-world" physics (no FTL or time travel).
First, let us examine the concept of "natural boundaries" as they exist on Earth.
I. Terrestrial Natural Boundaries
There are three basic sorts of natural boundaries: obstructive, logistical and cultural.
A. Obstructive Natural Boundaries
An obstructive natural boundary is the simplest sort. It is essentially a terrain feature that is difficult to pass; hence it renders defense along its line relatively easy.
Suppose that two countries, Alphia and Betasan, are separated by a mountain range which is crossable only by means of narrow passes. It is fairly easy for either country to build fortresses on its side of the passes, making it very difficult for an enemy army to force a passage. If both Powers be roughly equivalent in strength and competence, then wars between the two Powers will tend to be long, drawn-out and indecisive; for this reason they will be rare and unlikely to result in the capture of much territory. Hence, the border is likely to run along that mountain range for a very long time, until one or both of these countries collapse.
A real-world example of this is the Pyrenees between France and Spain: even though France and Spain fought each other several times between the 16th and 19th centuries, only during periods when one country or the other was exceptionally strong or weak were such wars fought to decisions on the enemy's home territories, and the diplomatic tendency was for the border to return to this natural boundary. One reason for this diplomatic tendency was that both sides were aware that the condition of strength or weakness on one or the other side was transient, and a border not drawn along this natural frontier would be unstable: it would simply lead to a war of reconquest in which the former owner of the territory would have all the advantages, for reasons which will be discussed under "logistical" and "cultural" natural boundaries.
Mountain ranges are obvious obstructive boundaries, but any sort of terrain which is difficult to cross can serve the same function. For instance a large river may only be crossable without boats at a few fords or bridges; a dense forest or wide swamp may have only limited roads along which supplies can be hauled, and of course operations across oceans are limited by the availability of sea transport, so coastlines are also natural boundaries.
Even today, where natural boundaries have to some extent been surmounted by the ability to build railroads, superhighways, modern seaports and airports, a surprising number of countries have coterminous natural and political boundaries. This is in part due to the fact that these boundaries may have been set before the Steam and Electrical Ages; in part due to the effects of logistical or cultural boundaries.
B. Logistical Natural Boundaries
A logistical natural boundary exists where a transport system, allowing easy movement of men and supplies, reaches its limits. Everything within the range of this transport system is relatively easy to defend, because armies may readily maneuver using this transportation. Beyond the range of the transport system, amies find it difficult to operate due to the need to drag large supply trains along with them: movement is slow and operations endangered by the risk of supply exhaustion.
The (literally) classical example of this is afforded by the Roman Empire. The Roman Empire had a basic transportation system afforded by oceanic shipping and riverboats, which it extended overland through the construction of the Roman roads. Because the Romans lacked steam locomotives and motor lorries, land movement of any great quantity of supplies had to be accomplished through beasts of burden.
But beasts of burden need fodder themselves in order to work, meaning that long-distance supply trains had to haul a lot of fodder just to fuel the beasts which were pulling their wagons. If the journey was too long -- beyond around 50-200 miles depending upon roads and terrain -- the supply trains would have to carry so much fodder that they could not also transport any significant amount of supplies, and hence this rendered military operations beyond this range both dangerous and impractical. (Steam trains and motor trucks have similar problems, but need much less coal or petrol per ton-mile than beasts of burden need food, hence motorized supply trains allow much longer supply lines).
In consequence, the Roman Empire basically followed the contour of the territories within a 50-200 mile range of the oceans and navigable rivers, save where deliberate strategic decisions were made to use a river itself as a natural boundary (as was the case at times on the Rhine and Danube). Indeed, sometimes within ostensibly Roman provinces, those areas far from the coasts and rivers would often be half-wild, with brigands and surviving semi-independent barbarians, because it was too much trouble for the Romans to police such regions save during prolonged periods of peace.
Obstructive and logistical natural boundaries can and often do coincide. For instance, the aforementioned Pyrenees are also a logistical natural boundary to any pre-Steam Age army, because it is difficult for supply trains to move through the passes. (In winter it's almost impossible). A French army trying to operate in Spain, unless it was able to get supplied by sea, might starve -- as the French discovered to their sorrow in the Peninsular War campaigns of the Napoleonic Wars.
Deserts are more logistical than obstructive boundaries. Most deserts are flat and many are relatively hard-surfaced: good walking terrain. But they are also hot by day, cold by night, and lack much in the way of available food and water. An army wishing to operate in the desert must bring its own water, which means it needs a larger than usual supply train, and the beasts pulling the wagons of the supply train also need to carry their own water.
Before the domestication of the camel, large-scale trans-desert military operations were almost impossible, and even afterward they were difficult. This is why the writ of the Pharoahs did not run too far from the Nile and its nearby oases such as Lake Fayum, even in eras when Egypt controlled major parts of the Levant. The Roman Empire was never able to expand very far south into Africa, save along the Nile, and the many wars between Rome and Parthia were usually rendered indecisive by the difficulty of crossing the Syrian deserts. Even after the development of motorized transportation, desert warfare was difficult: the North African campaigns fought between 1940 and 1943 rarely saw signficant forces move more than 100 or so miles inland from the coast, and operations were very much dependent upon the availability of seaborne supply and the control of good ports at which to land them.
When a good natural boundary is found and reached, a society may long defend these boundaries, and the people living within this boundary may develop a shared culture. This leads to the last major type of natural boundary.
C. Cultural Natural Boundaries
All things being equal, people would rather be governed by leaders and administrators of their "own kind" -- people who speak the same language, worship the same gods, and have similar philosophies of justice and standards of decorum. People will accept oppression from native tyrants far more readily than they would accept the same oppression from foreign ones, and the rule of even enlightened foreigners is often resented.
As I mentioned, since obstructive and logistical natural boundaries tend to be stable, the people within them will often be under the same rule for long periods of time. Over the generations, these people will assimilate with their masters and with each other, until they share a common culture: a culture that they perceive as very distinct from those people within or on the other side of natural boundaries.
This greatly increases the defensive power of the natural boundaries. Should an invader from outside the culture-group manage to surmount the natural boundaries and seize some or all of the territory within them, the resistance of the natives may be bitter and protracted. It may take centuries for the natives to accept and assimilate to the new rulers, and until then the governors will face a constant threat of revolt. If the imperial power faces other dangers -- internal or external -- at the same time, the threat may become an actuality, and the revolt may even prove successful.
This threat is far greater if the invader has conquered only some of the territories within the natural boundary system. As long as independent territories and armies of the native culture are still extant, such will serve as a rallying-point for the rebellious subjects of the empire, and such independent territories will be especially quick to strike at the empire when the empire suffers a time of weakness.
An excellent example of this is afforded by the fates of the European empires in the East Indies during and immediately after World War II. France and the Netherlands both lost the early phase of World War II to the Germans, which also weakened their now-unsupported colonial forces in Indochina and Indonesia. Britain remained undefeated, but could not afford their Malaysian colonies much support when the British were fighting for their lives at home.
The Japanese took advantage of their weakness in 1940 and 1941, invading first Vietnam and then Malaysia and Indonesia. Though the Japanese were thrown out of these territories by the victorious Allies in 1945, the postwar West was badly-drained by the economic, human and moral costs of the war, and the natives rebelled against their foreign masters. The Dutch were thrown out of Indonesia by 1950; the French from Indochina by 1954.
Even the British left Malaysia in 1957, though in part because Britain was less damaged than the other two colonial Powers by World War II, and in part because British rule had been more humane, the British were not forced out by the Malaysian Insurgency but instead remained friendly with the regimes of Malaysia and Singapore. Note that things might have been less friendly had the British insisted on holding on to power.
Cultural natural boundaries are subtler than the other two kinds of natural boundaries, but they can be very strong. Cultural areas can remain effectively unified even when they are formally governed by different regimes: note the strength of the Anglo-American Alliance from 1917 to the present date, or the friendship between all the nations of the Anglosphere (America, Britain, Canada, Australia and New Zealand). It is very difficult to engage in the long-term subjugation of an alien cultural group, unless one's military superiority and political determination both be great and long-lasting.
Now let us extend these basic concepts into the recent past, present, and future.
II. Spatial Natural Boundaries
Space is mostly empty and relatively invariant, so one might think that there are no obstructive boundaries in space. There are in fact two very large ones: gravity, and distance.
The "terrain" of space is formed by massive objects creating gravity wells, in orbit around other massive objects creating gravity wells. A force deep in a gravity well must expend energy to escape the gravity well; likewise, a force orbiting a gravity well must expend energy to counter the velocity of its orbit and drop into the gravity well -- save for missiles, it may also have to expend energy to avoid crashing into the massive object at high speed.
Distance is obstructive primarily because it provides time for defenders to destroy incoming missiles. Since missiles are potentially the most destructive of weapons, and since stealth is difficult in open space, the longer a missile has to travel before impact the greater the likelihood that it will be destroyed by defenders' weapons, particularly energy weapons. Given computers at least as advanced as those existing today, the effective accuracy of energy weapons asymptotically approaches 100% assuming a non-evading target, and evasion is essentially impossible at ranges below around 0.1 LS (about 30 thousand kilometers).
Secondarily, distance provides defenders more time to destroy incoming warships: the more so because energy weapons cannot quickly destroy a well-protected target but can slowly destroy such a target through overheating. The main defense against overheating is a heat sink, but heat sinks are ablative defenses (the heat must eventually be gotten rid of, mainly through dumping into hot volatiles which are then ejected), and given enough time the target will exhaust its supplies of coolant and be reduced to the much less effective means of heat loss by radiation.
Finally, at interplanetary or greater ranges, distance causes a communications lag which fatally lengthens the command and control loop. Thus, assuming no FTL couriers or radios, actual sapient minds must be present to do more than launch a long-range raid against any world not in the same lunar system.
Space is mostly analogous to a desert. Nothing save a lack of energy and reaction mass, or the existence of relatively strong gravity wells, prevents maneuver onto any desired vector. However, like a desert, everything (save for small amounts of energy obtainable by the reception of radiation) that one requires for supply must be brought along on the spaceship itself.
This means that distance is also a logistical barrier. If I wish to conduct military operations on Mars, and it takes my spaceship six months to get to Mars, then I must carry six months' worth of supplies for my ship, crew and equipment. Supply requirements may be reduced by means of recycling (utterly essential for rapidly-exhausted supplies such as air and water) and fabrication shops aboard the spaceship itself, but recycling and fabrication imply the provision of specialized equipment which also takes up mass and requires energy to operate.
One very important type of supply is reaction mass. This is distinct from "fuel" in all but chemical rockets: power comes from a fission or fusion reactor, or from chemical, nuclear or antimatter fuel cells, but this power cannot directly drive the ship. Instead, the power is used to fling out mass or energy, which then drives the ship through Newtonian reaction. This is a "reaction drive," and so far it is the only sort of space drive which we know how to build or are even certain is theoretically possible.
In space, there is neglible friction on a human timescale, so the important question of maneuver is not "speed" but "acceleration" and the resultant "velocity." Velocity is a vector rather than scalar force, which means that it is a magnitude PLUS a direction. One is not merely flying at, say, 20 km/sec, one is flying at 20 km/sec on a particular spherical bearing.
Changing one's vector is called "delta vee" (change in velocity), and any object with a certain reaction mass and exhaust velocity has a certain delta vee, and can be rated by that delta vee. When an object has exhausted its reaction mass, it cannot change its velocity save by very slow means relying on the weak force of the solar and other particle winds ("space sailing") and by the clever use of gravity and the tiny amount of friction which does exist in the (semi-) void. Hence, reaction mass (and delta vee) are a "supply" which is depleted by operations.
Worlds, from the greatest gas giant down to the tiniest asteroid, are potential supply sources. Whether this potential is realized is based on the resources and facilities present on the world and on the spacecraft. Gravity wells have a complex relationship here: the deeper the gravity well, the easier it is in terms of time and delta vee to achieve orbit; but the harder it is in terms of time and delta vee to leave orbit.
Acceleration capability is important in this regard. An ion-driven spaceship with a maximum acceleration of 0.005 G would take a lot of time to achieve orbit around a small world or leave orbit around a large one; a fusion-driven spaceship with a maximum acceleration of 5 G would have little problem landing or taking off (let alone acheiving or leaving orbit) from even the largest planet.
Local conditions may create logistical boundaries, because spacecraft may require special equipment to operate under particular local conditions. The ones we are most familiar with is atmosphere and gravity: a spaceship designed for operation in vacuum and under very low-g acceleration will not be able to take off or land from a world with a signficant atmosphere or gravity, and will have to conduct operations on its surface by means of shuttles or landers which have better engines or streamlining.
Another local condition is radiation, or the lack of radiation. A spaceship designed to operate closer to Jupiter than Callisto would need strong electromagnetic shielding, stronger than most spaceships might have. Consequently, the magnetosphere of Jupiter would be an obstructive barrier to ships without this shielding, and a logistical barrier in that ships would have to be equipped with this shielding to so operate. Conversely, a ship making extensive use of solar power would be useless in the Outer System, where solar radiation and hence the energies so obtainable would be feeble.
Distance, which we've already discussed as an obstructive boundary, is also a logistical one. If a force from the Solar System wishes to conduct military operations in the Alpha Centauri System, its communications with the Solar System will suffer an 8.72 year round-trip time lag, which means that even the fastest possible supply shipment (lasered energy) could not possibly be changed any more rapidly. This means that very long-range operations would have to be well-planned in advance, and could not be easily reinforced if they encountered unexpected difficulties.
Here we should mention receptive drive and energy systems, or to speak more familiarly, catapults and powerbeams. A spaceship does not need to carry all the fuel and reaction mass that it would normally if it can get some of its power through beamed transmission from bases, or some of its delta vee through launcher/catcher catapult systems. Several frequencies and kinds of electromagnetism (particularly masers) work well for beamed power transmission; options for catapults include electromagnetic rails, ion projectors, and lasers. Counterintuitively, a catapult "launcher" can also serve as a "catcher" -- in other words, a ship or cargo can be not only accelerated but also decelerated (most efficiently by a catapult at the destination, but it is actually also possible to do it from the point of origin using large reflector and receptor arrays deployed by the ship itself.
This mention belongs here because any such use of receptive systems requires coordination with the transmission facilities. This limits the flexibility of their use at ranges which impose signficant two-way communications lags. For instance, if a fleet from Earth is atempting to operate against Mars, and it is getting most of its energy by maser beam from Earth, it cannot engage in unplanned maneuvers (such as evasion or other combat events might require) or it will maneuver right out of its power beams. Likewise, any delta vee obtained from external sources cannot be used for unplanned maneuvers; in the case of catcher systems, the catchers themselves might be at hazard from enemy action. Hence, receptive systems are much more useful for civilian than military purposes.
Generally speaking, the maximum range of exploration will always be greater than the maximum range of colonization, and the maximum range of routine transportation less than the maximum range of colonization. This means that we will generally be able to explore farther than we can colonize, and colonize farther than we can engage in large-scale operations.
This is a situation quite familiar to terrestrial history, and its implication for the creation of cultural natural boundaries is that there will be a time period between the initial colonization of a world and the development of large-scale transportation infrastructure between the homeworld and colony. During this period, there will be signficant obstructive and logistical boundaries between the homeworld and colony.
Thus, the colony will be able to develop a culture distinct from the homeworld, and if the colony secedes from the homeworld, will enjoy significant defensive advantages should the homeworld dispute its independence. These defensive advantages will be the greater, of course, in a war fought long after independence, because a long-independent nation will have a more developed economy and military establishment.
Obviously, it takes time to establish a culture firmly enough, and to let its population grow through immigration or natural increase to the point that its cultural solidarity becomes a natural boundary. This is unlikely to happen, during the first expansion outward from the Earth, regarding Earth-Orbital or Lunar colonies, because the technological advances permitting large-scale military operations will likely come rapidly enough on the heels of those permitting colonization, and the vast existing population on the Earth will be very concerned about dominating Earth-Orbit, or Earth's only natural satellite. But it is very likely to happen regarding other worlds of the Solar System and beyond, because as we go farther and farther out, the number of potential targets for colonization increase to (literally) astronomical quantities.
These cultural natural boundaries will be important as a source of resistance to divisio et imperio ("divide and rule") invasion strategies: which is to say, obtaining a surface-head on and hence circumventing the other natural boundaries of obstructive distance and logistical supply support which act to defend a world against invasion from without. United worlds will be conquered from time to time, owing to the fact that an attacking multi-world empire may be able to concentrate vastly greater force to bear on a single planet than that planet has with which to defend itself, but the defender will enjoy a tremendous force-multiplier advantage, as long as he can remain culturally and politically united.
D. Technological Progress
The main instability inherent in any such system of natural boundaries is the advance of technology. Gravity wells that were obstructively-deep and distances logistically-wide to spaceships with one kind of drive may be conveniently-shallow and narrow to spacecraft equipped with later and more powerful engines. Logistical problems which crippled operations in one century might be solved by improved resource processing and product fabrication devices. Hostile local conditions that required massive and expensive protective systems at one level of technology might become mere nuisances to the improved protective systems that are later devised.
We have seen this repeatedly happen in terrestrial history. For instance, the very same ocean which prevented Great Britain from holding on to her North American colonies in the late 18th century could not prevent the United States of America from liberating Western Europe in the mid-20th. The ocean was no narrower, but what had been a major logistical barrier to sailing ships displacing hundreds of tons was only a minor one to steamships displacing thousands of tons. We now have routine tourist transportation to coastal Antarctica: to many of the same places which a mere century ago were reachable only at great danger and suffering.
In general, the key technologies which will overcome natural spatial boundaries are powerplants and engines, followed by life support and fabrication.
1. Powerplants and Engines
These two technologies are related, because greater energy-density of power system roughly correlates with greater exhaust velocity of engine, and hence the specific impulse generated by each kilogram of reaction mass (the relevant equation being f=m(v squared) where f = force, m=reaction mass, and v=exhaust velocity). As this demonstrates, increasing the exhaust velocity has much more effect on the production of thrust than does merely dumping more reaction mass: reaction mass affects thrust force linearly while exhaust velocity affects it geometrically.
Generally speaking, the simplest kind of rocket is a chemical rocket, in which fuel and reaction mass are one and the same. Chemical rockets allow routine access to orbital space but allow access to inter-lunar space only with difficulty and inner inter-planetary space (Mercury to the Asteroid Belt) with extreme difficulty. Next comes nuclear fission and thermal or ion drives, which allow routine access to inter-lunar space, inner inter-planetary space with difficulty and outer inter-planetary space (Jupiter through Neptune) with extreme difficulty. Then nuclear fusion and plasma or fusion drives (a fusion drive is a plasma drive with a fusion afterburner), which allow routine access to inner inter-planetary space, access to outer interplanetary-space with difficulty, and far outer interplanetary-space (the Kuiper Belt and Oort Cloud) with extreme difficulty. Finally come antimatter-powered photon rockets, which allow routine access to outer interplanetary space, far outer interplanetary space with difficulty, and the stars within about a dozen light-years with extreme difficulty.
There may be power systems and drives beyond that, without totally leaving the confines of known or almost-known physics. For instance, energy might be drawn from the structure of spacetime ("Zero Point" energy) or stored in stretched gluons (by teasing and holding apart the quarks in subatomic particles), or fuel or reaction mass might be scooped from the interstellar medium through the use of lasers to ionize and electromagnetic fields to draw in interstellar dust and gasses (the "Bussard Ramjet" principle). Considerable advances in physics and engineering would be needed to employ either method on a practical scale, but none of this would violate known physical law. Such improved systems might allow merely difficult or even routine access to the close stars.
Beyond this of course lies the possibility of FTL communication and travel, but since so much of the physics involved is incredibly speculative and I'm trying to stick as close as possible to "hard science," I will leave off at this point.
2. Life Support and Maintenance
Both are essentially concerned with mastery of chemical and mechanical processes, and neither involve any fundamental physical advances: they are purely issues of design and control of the systems. Hence they may be assumed to be developed further and further with time; yet the precise degree of their development at any particular time imposes the most basic logistical limits on military and other operations.
We already know what chemicals and biochemicals human beings need to survive. Basically, we need clean air and water, and wholesome foods. Providing such, on the scale of a settlement or even a large hab, would be a necessary but not theoretically very difficult task. The problem, of course, is that for transport operations, one must either shrink the recycling and production systems down to a portable mass, or simply accept that one will only be able to produce an incomplete set of the required biochemicals and supplement that set with irreplacable supplies, carried aboard the spaceships.
Likewise, we would know precisely what sorts of replacement parts might be required to keep functional the mechanisms of the spaceship and its equipment (including the life support system). As technology advances, the mass and other costs associated with the required highly-flexible machine shops reduce, and the more sophisticated the equipment which can be repaired onboard, rather than at a better-equipped base.
However, there is a subtler problem associated with both life support and maintenance, which may prevent logistics from ever being more science than art. That is that both ecologies (even the limited ecologies of spacecraft life support systems) and systems of mechanisms (such as those comprising the spacecraft) are complex systems which can succeed or fail in complex, synergistic, "emergent property" fashions. Consequently, both sapients of skill and emergency backup systems and supplies will always be needed, albiet to a diminishing extent as the technologies progress. And there will always be the occasional disaster, avoidable or unavoidable, to remind sapients that traveling across long distances is inherently hazardous, though also to a diminishing extent as progress marches on.
III. Levels of Boundary
The Great Powers of the Earth have been an effectively sub-orbital civilization since the 1950's, when we developed long-range rockets capable of flying accurately for thousands of miles. By putting nuclear warheads on these rockets, and by putting these rockets on submarines, surface ships and heavy bombers, we effectively rendered all existing terrestrial obstructive natural boundaries irrelevant at least where unlimited strategic bombardment was concerned.
In consequence, the whole world of Earth has now become effectively part of the same natural boundary system. Any Earthly Great Power may carry out strategic bombardment against any other Earthly Great Power, and the only military bar to such bombardment is an active defense: the deployment and use of anti-missile missiles or defensive energy weapons: and such weapons are not significantly enhanced by the presence of any natural boundaries save (to a limited extent) large oceans.
Logistical natural boundaries have been severely weakened by the wide deployment of railroads, superhighways and large steamships. Where such boundaries are still somewhat effective, it is through defensive action: for instance, interdiction strikes on enemy rail and road nets, or submarine and air strikes against enemy transport shipping. They are becoming increasingly ineffective however as transportation becomes cheaper and more versatile. It is much harder to interdict motor lorries than it is steam trains, and harder still to interdict transport helicopters.
Cultural natural boundaries are still strong, because they take generations to drop. Widespread global communications are slowly breaking down such boundaries: it would already be politically very difficult for two Western Powers to go to war against one another, and in such a case both parties would face extreme political pressure to conduct a war in the most humane possible fashion (for example, the Falklands War, in which both sides accepted surrenders and treated captured military and civilian personnel very correctly). Most vicious wars are now either fought by non-Western Powers or across strong cultural boundaries (and the latter, of course, tend to be indecisive for precisely that reason).
The obvious long-term trend for military operations on any one world in a state of sub-orbital civilization is for them to either lead to general devastation (as strategic bombardment becomes too easy) or general unification (whether through peaceful federation or protracted conquest). Eventually, any world which survives its disunited stage will become culturally and politically unified, and then its own ionosphere becomes a "natural boundary."
In the not too distant future, sub-orbital capabilities may also be used to reduce logistical barriers to the very rapid deployment of military forces. There is no theoretical reason why hypersonic sub-orbital rockets might not be used as troop transports, though the economics of such rocketry will mean that they will at first only be used for the insertion of very small specialist forces, such as commandoes and various kinds of scouts and engineers. Eventually, most intercontinental air transport might be superseded by suborbital passenger rockets or (magnetic or other) catapult-launched capsules.
Before this happens, though, the nature of the technologies involved tends to lead to such a world ascending to the next stage of transporation capacity.
In our history, it took less than two decades to pass from the first sub-orbital rocketry to the first orbital rocketry. However a case can be made that we did not really become an "orbital" civilization until the deployment of the first manned orbital space stations in the 1970's.
Because there are no natural objects in close orbit around the Earth, becoming an orbital civilization did not mean any major changes to military operations in terms of new terrains to colonize and hold. This might not be the case for all orbital civilizations: for instance, if an orbital civilization had evolved on Mars, it might have progressed rapidly from sub-orbital rocketry to planting actual settlements on Deimos and Phobos, both of which are in fairly close orbit of their planet. We may also in the future choose to build large permanent close-orbital stations (though there are some disadvantages to such siting, due to the need to periodically trim their orbits to counteract the effects of high-atmospheric drag).
Instead, the main effects on military operations of becoming an orbital civilization were to vastly improve the power of reconaissance for any Power with space superiority, or even parity, with its foes. The development of reconnaissance satellites greatly reduces the ability of distance or obscuring terrain to hide forces, which in turn weakens natural boundaries.
An artificial space settlement in close orbit around a world would be highly-vulnerable to military action from the world's surface. It might be able to defend itself from enemy missiles, because the missiles would have to cross a large open space to reach the settlement, and during that time the settlement would be able to engage the missiles with their own anti-missile missiles or energy weapons, but enemy energy weapons based on or near the world's surface could bombard the space structure with extreme accuracy, burning precisely into its sensors, weapons and power systems. Even an armored space settlement would eventually succumb to the power of enemy beams which would have a whole world's power grid on which to draw, and which would be commanded from bunkers miles-deep in the world's crust.
The defense of a moon in such an orbit, however, might prove more practical. Though enemies on a world's surface could still snipe with energy cannons at exterior targets with extreme accuracy, the sheer volume and mass of a moon -- even a small moon like Phobos or Deimos -- would enable the defenders to make a literal "defense in depth," meaning that they could locate their command centers, arsenals, barracks, depots, powerplants and even populations miles deep beneath the surface, rendering bombardment with energy weapons long and difficult propositions.
The defenders' own energy weapons would be able to draw on the gigantic stationary powerplants which one might site on an actual astronomical body. They could be located in tunnel systems, spending most of their time safely hidden in the moon's depths, scooting up to fire, and then once again retreating to avoid counterbattery bombardment. Sensors might be similarly concealed, and as long as any sensors remained operational, the moon could put forth such a volume of defensive fire as to render either missile bombardment or orbital assault impractical.
An energy weapon, if it fails to immediately-damage a target through thermal explosion, inflicts harm through the cumulative conversion of its energy to heat. This means that a major defense against energy weapons is to pump the heat away from the point of bombardment to a heat sink: a large relatively cold mass which can absorb the heat and, eventually, vent the heat into space either through a radiator array or by the emission of hot gas. An astronomical body has a clear advantage over a ship or all but the largest conceivable habs in this regard, in that it is possible to create immense storage tanks for volatiles that can be used for such heat dissipation systems.
C. Lunar Systems
The next step up is the "lunar system," which comprises a planet and all its associated moons. The vast majority of planets have moons (in our own system, Mercury and Venus are the only moonless terrestrial worlds): all known Solar gas giants have multiple moons and even some dwarf planets, such as Pluto, Eris and Haumea, have lunar systems).
The reasons why a lunar system is a natural unit is that a lunar system -- even the biggest ones such as that of Jupiter -- is small enough that (a) even chemical rockets can (with difficulty) travel about such a system and (b) energy weapons fire can be delivered more or less accurately against even maneuverable targets elsewhere in the same lunar system. Consequently, once one colonizes any object in such a system, one can readily colonize the other such objects, and a State in one part of such a system can readily project power against other parts of the same system. The imaginary border of a lunar system is therefore an "obstructive" (defense-moat) and "logistical" (transport-net) natural boundary, and in time lunar systems are also likely to become cultural natural boundaries.
As discussed under "Orbital" operations, any moon (even a moonlet such as Deimos or Phobos) is large enough that its possession gives the possessor a significant defensive force multiplier. This might enable a State controlling one moon to retain its independence from another State -- even a stronger one -- based on the planet or on another moon in the same system.
What one then gets would be a system of "moon-states" which would be in many ways analogous to that of the "city-states" common in the history of Earthly regions divided by land obstacles but united by greater surrounding natural boundaries, such as Greece or Italy. This system is unstable in the sense that advancing technology will inexorably increase the capability of interlunar transport and the firepower and defense of warships, eventually leading to a situation in which the interlunar boundaries must collapse either to confederation or conquest; but depending how long this process takes to complete, cultural natural boundaries might appear and delay the resultant unification.
This is not very likely to happen, at least initially, in the Terrestrial lunar system (Terra and Luna) because the Earth starts with an immense population and industrial base (billions) while the Moon will at least at first only have colonies measurable in the dozens, hundreds, or thousands, with very little industry in absolute terms by comparison with that of the Earth (though probably much more industrial output per capita). Long before the Lunar population has increased to the point of being able to credibly defend a whole world, Terra will have enough atomic-powered rockets to transport enough ordnance and manpower to overwhelm any likely Lunar defense. (This is a shame, as one of my favorite science-fiction novels growing up was Robert A. Heinlein's The Moon Is a Harsh Mistress).
In the case of Mars, the primary world is an obviously-attractive target for colonization and economic exploitation, while the two moons are small and uninteresting. Hence, the Martian population will quickly eclipse that of its moons, which will probably function primarily as spaceports and fortresses serving the needs of the mainworld's society. High population turnovers are likely, with consequently weak cultural formation on the moons. Now and then an enemy or adventurer may succeed in seizing one or both moons from the Martian., but this is likely to be merely short episodes in Martian history, assuming a unified Martian planetary government.
Lunar systems, however, will be especially important to the histories of our four gas giants. Jupiter has four large moons (Callisto, Ganymede, Europa and Io) each of which is likely to be of interest for slightly different reasons as a target for colonization (Callisto, outside the magnetosphere, is an obvious spaceport for the Jovian system; Ganymede has seas each of which may have different ecosystems and which are sources of volatiles; Europa has a gigantic planetary ocean likely to boast life; and Io is dense and volcanically active, hence a good target for mining). Jupiter herself is a hostile environment for colonization, hence it is likely that the Jovian moons and especially the four big "Galilean" moons (so-called because Galileo first discovered them) will boast significant populations before Jupiter.
The Jovian System is also highly-defensible against attack from the rest of the Solar System for several reasons. First, the Jovian magnetosphere both provides energy to any Jovian civilization, which could also be used to power energy beams and defensive shields. Secondly, that same magnetosphere requires strong shielding to operate anywhere within Callisto's orbit (Callisto herself is outside its worst effects) and hence most interplanetary cruisers, especially in the early ages of interplanetary civilization, will probably not be able to participate in attacks into the Jovian system without extensive refitting. Finally, it is a large and diverse lunar system which would offer its occupants access to all kinds of metals and volatiles, hence well able to withstand blockade.
Though at first one would expect to see moon-states, eventually some sort of Jovian Lunar League or Federation seems a likely governmental type: the sooner if the Jovian System feels threatened by some external force. The natural boundaries around the Jovian System being very strong, such a League would be breakable only through application of overwhelmingly-greater force or through the disunity of its members.
The moons of this federation would probably take charge of efforts to colonize Jupiter, which would occur because Jupiter is a potentially-limitless source of energy and minerals. The colonization of Jupiter would be very difficult: even establishing floating settlements in the upper atmosphere would require operating in a 2.54 G field, which would practically require extensive use of cyborging, genetic engineering or robotics on the part of the colonists. As one goes deeper, pressures and temperatures climb, and the Jovian core is one of the most hostile places in the Solar System (far worse than the core of the Earth or the surface of the Sun). But it is also a treasure-trove: a vast silicon-nickel-iron-actinide body larger than the Earth, probably very active and hence containing immense concentrations of virtually any desired heavy element, so the rewards for building or breeding sapients capable of operating under these insanely-difficult conditions would also be great.
One problem that a Jovian Lunar Federation might face is that the natures, and hence perhaps the cultures, of the beings capable of operating under the extreme conditions of the Jovian depths would be very different from that of unmodified human beings or even humans modified to operate in the Jovian moons or in the Jovian upper atmosphere. Since the colonists of Jupiter (the "true Jovians") would eventually gain access to resources far greater than those obtainable on the Jovian moons, a long-term tendency might be for the Jovian moons and their inhabitants to fall under the sway of their own Jupiter colony or colonies.
The other gas giant lunar systems are smaller and less rich than Jupiter, but they too contain their prizes: also, they too would tend to see their moons colonized before the main bodies, and in the long run to see the power of the planetary colonists eclipse that of the lunar colonists. Saturn's lunar system, of course, is dominated by Titan which is large, fairly dense for an object in the Saturnian system, and immensely rich in volatiles including hydrocarbons. There are also several other moderate-sized moons, but it is likely that the lunar system would be dominated by the colonists of Titan.
In general, Saturn and its moons are fairly low-density objects and hence could be expected to be metal-poor; Saturn's rocky core is also small compared to Jupiter's, but even so is still larger and richer in absolute terms than a terrestrial planet. Thus, in the long run the whole Saturnian system might fall under the domination of its mainworld.
Uranus has several moderate-sized moons, of which the most important are Oberon, Titania, Ariel, Umbriel and Miranda. Oberon, the outermost, might be the system port, while the other moons are sufficiently large and diverse that a lunar-state system might emerge. The Neptunian lunar system, in contrast, is dominated by its biggest moon, Triton, which would almost certainly control the other moons as well. Both the Uranian and Neptunian systems would eventually wind up dominated by their mainworlds: either giant would be much easier to colonize than Jupiter, though such colonization would offer comparatively less rich rewards.
Pluto and Charon are very nearly a double dwarf planet: one might long see a situation where each world remained independent of the other. The other Solar dwarf planets have only tiny moons compared to their mainworlds: Eris has Dysnomia, Haumea has Namaka and Hi'iaka, and so far Ceres and Makemake appear to have none.
It must be emphasized here that all these lunar systems have immense defensive advantages against interplanetary attack. Approaches even with plasma or fusion drives would take weeks or days, during which the defenders could scourge the invading armadas with gigantic energy weapons powered by huge stationary fusion reactors. Until the invaders seized at least one moon in the system, they would have no safe base at which to rest and repair damage, and even then they would have to hold off counterattacks at merely interlunar distances, while invader reinforcements would have to travel across interplanetary distances, and worse would have to pass through the same sort of gauntlet of long-range defensive fire. Unless numerous and well-planned, such invasions could easily end in disasters for the invading forces.
This leads us to
The scale now expands from mere tens or hundreds of thousands, or at most millions of kilometers; to tens to hundreds of millions of kilometers: a two to four-fold increase in order of magnitude. This increase is conditioned not by arbitrary human definition but by fundamental astrography: there is a big gap in scale between lunar systems orbiting planets and planetary system orbiting suns.
It is perhaps best to explain this in terms of what military operations even an interplanetary civilization could not easily perform. It could not perform interplanetary bombardment with any great chance of success assuming even roughly equal force strengths (the defending lunar system could shoot down salvoes of missiles with great ease), until perhaps very fast kinetic or even relativistic kill vehicles were available (and perhaps not even then, as outlying space fortresses and large-scale electromagnetic shields powered by whole planetary magnetospheres might offer significant defense in depth against such attacks). It could not invade another lunar system save as part of a very large and well-coordinated military effort, and even then would be at serious risk of disaster. And travel between planets would always be more difficult, expensive and time-consuming than travel between parts of the same lunar system, with all this implies regarding the difficulty of timely reinforcement of an interplanetary invasion.
A natural unit is formed by the worlds of the Inner Solar System, which are mere tens of millions (as opposed to hundreds of millions or billions) of kilometers apart. Against light opposition (such as early colonies would put up against imperial forces from Earth), even nuclear fission powered ion-drive warcraft might suffice to maintain control; and nuclear fusion powered plasma-drive ships, especially if supported by catapult launcher-catcher systems, could make the entire Inner System a single economic and defensive zone from the point of view of any of the outer planets.
The obvious natural boundary is formed by the edge of the Main Belt, which offers plenty of mass and natural astronomical bodies on which to base the control centers for an early warning system able to detect and to some extent intercept attacks coming from beyond along the plane of the Solar ecliptic (the Belt would also be a good place in which to build a thinner early warning sphere to deal with possible attacks coming in above or below the plane of the ecliptic). One convenient aspect of the Main Belt is that many of its asteroids orbit as much as 20-30 degrees off the ecliptic, providing astronomical bodies on which to site control stations for detectors even off the ecliptic, reducing the thinly-covered zone to around 120 degrees to the "north" and "south," and this far off the orbits of all but the dwarf planets and thus requiring high-delta-vee courses to navigate (and high delta vee means more engine burns and hence easier long-range detection).
Because the Main Belt is large (over 6 AU in diameter and hence some 19 AU in circumference), this early warning sphere would of necessity be thinly crewed, with single stations acting as control and collection systems for small cheap arrays spread out over millions to tens of millions of km of the surface of the sphere. The largest such stations would be fortresses and bases for spacefleets, but most of the fleets themselves would be closer in to the Sun, in the middle or inner parts of the Belt or even near Luna or Mars. When an incoming threat was detected, depending upon the geometry of the situation, fleets would be vectored in to intercept the invader, a classic thin-shell mobile force system, similar to that employed by the Late Roman Empire.
The Trojans and Greeks, groups of asteroids roughly in and in orbital resonance with Jupiter, would have a different significane. Each group is sufficiently spread out to constitute its own interplanetary zone rather than a part of Jupiter's. They are of course easiest to reach from Jupiter's own orbit, and control of them would have strategic implications both for the defense of Jupiter and of the Belt. If Jupiter was controlled by the same polity which controlled the Belt, there would be defensive outposts in the Trojans and Greeks; if they were controlled by different polities, there might be conflict over the settlement of these bodies.
Jupiter, Saturn, Uranus and Neptune have sufficiently large lunar systems to be seen as interplanetary groups of their own. Also, each giant has accompanying asteroids (the Trojans and Greeks in the case of Jupiter, and some of the Centaurs in the case of Saturn, Uranus and Neptune) easily reachable from their lunar groups. In the case of Neptune, this includes some of the inner fringes of the Kuiper Belt, including the Pluto-Charon lunar group. Each giant lunar group would have strategic interests in its associated asteroids, which could otherwise be used as bases against their systems.
There are presumably groupings of dwarf planets and planetoids in the Kuiper Belt and Oort Cloud, but not very much is known about the former and almost nothing about the latter. Eris, Haumea and Makemake, the largest known such objects, presumably have many attendant smaller bodies. The distances out here are immense, and "attendants" might easily be farther from their primary worlds than the width of the whole Inner System.
The natural boundaries betwen parts of a Solar System (such as the Inner and Outer Systems, or the Outer System and Kuiper Belt) would be largely logistical rather obstructive. The main problem here is distance, and differing design philosophies for warships intended to operate in different parts of a star system. For instance, a warship operated by an Inner-System polity, and never expected to travel beyond Jupiter, would carry extra powerplant, weapons and armor at the expense of engines and cargo capacity: it might be very effective in direct combat but able to traverse billions of kilometers only with the aid of an entire fleet supply train; a warship operated by a polity centered around Pluto or Sedna might have relatively light weapons and armor, but would have top-notch detectors, powerful drives and capacious cargo holds to enable it to cross immense distances without external logistical support.
This difference would be exaggerated by differing resource availabilities. The terrestrial worlds of the Inner System have lots of metals and direct access to vast quantities of Solar energy, and hence could easily produce the super-alloys and exotic matter required for armor and weapons. The moons of the gas giants are much less dense, but might have access to the rich metal deposits of the gas giant cores. But the iceworlds of the Kuiper Belt and Oort Cloud are mostly frozen volatiles over silicate rocks: while they would have no problem obtaining energy from the abundant hydrogen in those volatiles, they would skimp on metals and exotic matter as much as possible, reserving them for detectors and engines: much of their structure and armor would consist of nanocarbon and nanosilicate plastics. Hence, available resources would further affect design choices.
A major limitation on military operations at such vast distances would be communications time lags. An Admiralty conducting operations at ranges of light-hours to light-days from their main headquarters would find it very impractical to micromanage their forces: realistically, they would have to assign general goals and trust to the officers at the front to intelligently work towards these objectives. Any Admiralty which insisted on such micromanagement would simply doom their own forces, as their enemies --operating closer to the enemy headquarters or with greater local control -- got inside their command and control loops and defeated them in detail.
This would greatly affect the military cultures of the polities involved. The officers of an inner-system polity would tend to refer problems up to higher command, deferring their own decisions to their superiors whenever possible. In contrast, the officers of an outer-system polity would be used to having to make important decisions on the spot, with their superiors only knowing about these decisions in retrospect. This difference might be further exaggerated by the likelihood that it would be those with higher levels of initiative and independent spirit who would tend to colonize the outer system in the first place.
Maneuvering over these distances would be highly impractical without at least plasma and preferably fusion drives; it is quite possible that by this point warships would be equipped with antimatter power storage and true photon rocket engines. Such drives would themselves be powerful weapons, but of course by then their true weapons would be far more effective.
It is at around this point that relativistic kill vehicles would start to be practical, because the power densities involved would allow the launching of reasonably large RKV's both from bases and from warships. The main defensive challenge imposed by an RKV is the great speed at which it travels (a signifant percentage of the speed of light) and the fact that, even after interception, the result is a mass-to-energy conversion explosion which leaves an expanding sphere of plasma still on roughly the same course at close to the speed of light. Consequently, any target defended against RKV's would require detection and defense in depth, and powerful energy shields in point defense, to ward off the potential damage inflicted by the near-C ionic and subatomic cloud of debris. The use of and defense against RKV's is a complex topic, worthy of an article in and of itself, and I will not go into it too far here.
The extreme outer parts of a star system would be the most susceptible to the development of independent cultures and hence cultural natural boundaries. This is because their communications with the inner parts of their own systems -- or other parts of their own Oort Cloud -- would be subject to great time lags, and would be closer to long-distance mail in the 18th or 19th century than to modern telecommunications. On the other hands, delayed posting protocols on message boards might allow meaningful dialogues even with such restrictions, and the quality of the messages might still be high with redundant signal techniques, so some sort of cultural unity or at least strong sympathy could be maintained within a given star system.
If a civilization is able to travel to the edge of the Oort Cloud, 1-2 LY from the Sun, it is almost half of the way toward having the capability for interstellar travel. This would require nuclear fusion or antimatter powered photon rockets, possibly augmented by electromagnetic ramscoops. Even at decent sublight speeds -- say around 0.5 C -- it would take 9-25 years to reach one of the closer stars, so starships would need to be large and have extensive life support and self-maintenance capabilities.
It is debatable whether or not the first starships would be launched as major prestige projects from the Inner System (Mercury is a particularly good place to build starships, because one can benefit from immense amounts of solar energy and a powerful slingshot effect at launch), or as a more incremental extension of inhabited areas from the outer Oort Cloud onto rogue worlds (terrestrial or dwarf planets drifting between the stars) or into the Oort Clouds of other star systems. In any of these cases, we would be talking about large, powerful long-duration transport vessels.
Once colonies had been founded on the worlds of other star systems, there would be a great tendency for their cultures to diverge from those of home system. For instance, a colony civlization in the Tau Ceti system would suffer a 12-year one-way and 24-year two-way comunications lag with the Solar System; it would react to the culturally-influential events of over a decade ago and might not react in the same ways as did the home or other colony systems.
This would make it extraordinarily-difficult for an interstellar government to exercise any meaningful control over the administration of such a colony, unless the control was very light and rare indeed. The best structure for an interstellar polity would be some sort of league or very loose federation: any attempt at a tighter regime would be very likely to provke revolts.
The drives required to make starflight practical would make unifying any particular star system very easy, as any vessel designed to be able to make decade-long voyages of 4-5 LY distance would encounter few difficulties in merely crossing a star system. To travel from edge to edge of our own Oort Cloud, for instance, would be a voyage of merely 4 LY, which is not much shorter than the 4.36 LY to the Alpha Centauri System.
At the same time, by making interstellar colonization practical, this technology would introduce an extreme likelihood of its civilization fissioning into many smaller civilizations, and protect those civlizations behind a truly tremendous natural boundary -- the vast gulfs, with all that this implies in terms of communications lags and logistical obstacles, between the stars.
Any sublight interstellar military operations would have to be very self-sufficient, able to at least secure system-heads and hold them for years or even decades before being reinforced, because communications lags would require such time to respond to changing situations and dispatch relief forces. Of course an operation could be planned in detail including follow-up fleets, but if each fleet -- even the reinforcing ones -- were not capable of securing a lodgement on its own, then this would open the distinct possibility of total disaster, with each fleet in turn flying into its own destruction. Interstellar wars might then acquire a World War One flavor, with immense sacrifices being made over long periods of time for small and incremental gains.
Relativistic Kill Vehicles might break the stalemate, particularly if they could fly at very close to the speed of light. There are however problems with very fast RKV's (in particular, their sensors and effectors might become damaged by collisions with interstellar matter) and my estimate is that the attacker does not enjoy anything like the total advantage envisioned by Pellegrino and Zebrowski (The Killing Star), at least against a systemwide civilization with equivalent technology and any time to prepare before the outbreak of war. RKV's also offer the distinct disadvantage that they can't conquer anything: they are purely weapons of bombardment, and if launched across interstellar distances rather than by starships arriving on the scene, their bombardment plans are severely inflexible. In extreme cases of distant launch and defense in depth, a defender might have time to move a planet out of such a salvo!
Logistical problems related to fuel are very severe, unless electromagnetic ramscoops become practical. This could lead to a situation in which a fleet arriving in an enemy or uninhabited system might be forced to construct a base simply to refuel, which might impose long timelags on interstellar operations in addition to the transit times. If this is the case, then RKV's become even less attractive as an aid to conquering a system, as the RKV's would destroy the very facilities which one would hope to use for resupply.
My guess is that RKV's will be used, but rarely as the sole weapon of interstellar war, and more often with salvoes timed to support and be controlled by the sapients aboard an invading fleet. Such salvoes, controlled by people in system rather than watching the battle from many light-years away, might even be tactically-employed against enemy forts and fleets, instead of merely serving as a means of bombarding inhabited worlds. Combined-arms operations of this sort might prove the most effective means of interstellar combat. Large and powerful star cruisers might also be able to launch their own RKV's as anti-shipping missiles, much as modern naval warships employ SSM's.
Defenses against such will be very important. United systems will deploy early warning surfaces at the outer edges of their own Oort Clouds, with thickening detection and weapon stations and fortresses as one moves further in. Approaching the Outer Systems themselves, there will be major fleet bases with large fleets of warships ready to move to intercept incoming RKV salvoes, and batteries of very long-range anti-RKV missiles. Since an RKV could be killed by putting so much as a dust mote onto a direct intercept course, where detection is early the defense will enjoy the advantage. In the Inner System, vast electromagnetic arrays will power "sunbeams" (thank you E. E. "Doc" Smith for this term) which will redirect large fractions of total stellar outputs into super-powerful defensive beams able to scour wide cones of space clear of all incoming enemy weapons.
Nevertheless, the distance between the stars is something of an ultimate natural boundary, from the point-of-view of known physics. One might envision interstellar wars between rival federations or empires, in which interior lines and the greater defensibility of continguous frontiers became important: there is no inherent reason why the early-warning lines could not be extended between adjacent star systems, making flanking such defenses difficult for an attacker. Beyond a certain size, though, such wars would be unmanagable absent FTL radioes, and would tend to degenerate into mere interstellar anarchy.
So we leave our overview of future spacefaring civilizations and their extents with system-states and small interstellar empires. Is anything vaster possible?
Remember this: much of physics is still unknown. And the future may be far stranger than I -- or anyone -- can possibly imagine.