Can't help it! Have to add my two cents, pedantically.
Hint: How do we spell "SUPERSATURATION"? How about "NUCLEATION SITES"?
No matter how the CO2 is added to a soft drink, it is obvious that there is more CO2 in a soft drink than would be in equilibrium with the small content of CO2 in atmospheric air since a soft drink container is pressurized, at least slightly, when you buy it. The vapor space of the container has a fairly high content of CO2; still not enough to cause any harm.
The amount of total "CO2" dissolved into the liquid is primarily determined by the following equilibriums:
1) Physical dissolution of CO2 gas
CO2(g) <--> CO2(aq) math: [CO2(aq)] = Kg PCO2
i.e. the concentration, [] indicating mole per liter of CO2 dissolved into the liquid where it is part of the aqueous (water) solution ( thus CO2(aq) ) is related to the partial pressure of the CO2 gas, PCO2, and the constant relating these quantities, Kg, (or 1/Kg', depending upon definition) is Henry's Law constant which varies with temperature, somewhat with concentration of other components of the solution, and slightly with pressure (usually neglect pressure variation, particularly up to about 5 atmospheres pure CO2). At total pressures generated by small pumps, forget about the total pressure making a difference in this equilibrium. The total pressure is sum of the partial pressures of the gas components so adding air (~79% nitrogen and ~21% oxygen and only about 0.03% CO2) pressurizes the container by increasing the partial pressures of nitrogen and oxygen. Okay, there is a small increase of CO2 partial pressure but it is very small. Okay, also there is some increase of dissolved nitrogen and dissolved oxygen due to the increased pressure (same equilibrium principles but smaller constants); the dissolved N2 and O2 will pop out when the pressure is released but it is small compared to CO2. There is a another reason for increased pressure (with an air pump) to help maintain "freshness" (carbonation), see below, but it is the CO2 partial pressure which governs in the dissolved gas equilibrium not the total pressure (to any appreciable extent). However, CO2 also interacts chemically with water, which nitrogen and oxygen do not. Hence, more equilibriums...
2) Hydrolysis
CO2(aq) + H2O <--> H2CO3
math: [H2CO3] = Kh [CO2(aq)]
Carbonic acid ( H2CO3 ) is formed which is actually not that weak as acids go meaning it dissociates fairly easily. Hence ...
3) Dissociation
H2CO3 <--> HCO3- + H+
or H2CO3 + H2O <--> HCO3- + H3O+
math: [HCO3-][H+] = Ka [H2CO3]
These latter equilibriums increase the total amount of CO2 that is absorbed into a water solution. With the right data for these equations, one can calculate the total amount of CO2 dissolved in the solution but ONLY WHEN EQUILIBRIUM is reached which can take a long time. This might be done for the original carbonation process (e.g. supposing the liquid is highly pressurized in a stirred vessel with a pure CO2 gas before bottling) and calculations would give the residual amount of CO2 in a soft drink that has gone totally flat and is in equilibrium with atmospheric air pressure.
When a soft drink is opened, the partial pressure of CO2 above the liquid is reduced but it doesn't explode. Yes, shaking before opening will cause a lot of initial "foaming" and you can feel a greater pressure on the wall of a plastic container that has been shaken but this is the result of the liquid already being supersaturated with CO2 and carbon dioxide seems to supersaturate water solution quite easily. Other gases can also supersaturate but not as much as CO2; don't forget about those other equilibriums which also take up CO2 into the solution.
Supersaturation means that the actual concentration of the CO2 in the solution is higher than the equilibrium concentration and it is a meta-stable condition like an egg balanced on a sharp point. A shock like a sudden temperature change (contact with ice), or a shake will induce a change towards the more stable condition; CO2 gas will come out of the liquid.
In a closed bottle, shaking causes some gas to come out and the partial pressure of CO2 increases so the total pressure increases as well; it doesn't take much CO2 release to raise the partial pressure enough in a small space to attain a condition near equilibrium. If the bottle is opened at this point, the agitated contents will spew forth since the shock is not yet dissipated, but if the bottle is left standing for a long while, things settle down, the higher pressure dissipates through the cap seal (and somewhat by permeation through the plastic) and the solution becomes supersaturated again as the pressure falls, there is not a lot of CO2 lost but the drink will become flat a bit sooner than if the bottle was not shaken.
Even without any shocks, the CO2 will continue to be lost as the supersaturated condition still slowly works toward equilibrium. As noted in the other replies, this is faster in a plastic bottle than in a glass bottle due to permeation loss. Smaller plastic containers do have more permeation loss due to a larger specific area (area per volume). In a more impermeable glass bottle, the loss is more likely around the cap seals which aren't perfect.
When a container is opened, air mixes with the gas in the vapor space of the container, CO2 content/partial pressure decreases and there is a greater driving force for CO2 to come out of solution. Yet, leaving the container alone, it will still take quite some time for equilibrium to be reached and the soft drink to be fully flat. First the dissolved CO2 near the surface comes out, as the concentration of dissolved CO2 decreases, the other chemical equilibriums shift towards dissolved CO2, more dissolved CO2 can come out of solution. As the concentrations near the surface are depleted, more of the CO2 components from lower depths diffuse upwards, etc. So it takes awhile.
Pouring the liquid into another container, even very slowly, will set up slow currents that help to keep the concentration even throughout the whole volume and the CO2 will come out faster. Especially in the container receiving the liquid while the remaining liquid in the original container still seems to retain its carbonation longer even though it probably has similar currents set up due to the pouring action (okay, probably somewhat to a lesser extent). However, the major factor for the faster rate of CO2 loss in the receiving container is due to the exposure of the carbonated liquid to a new set of nucleation sites.
The CO2 gas coming out of solution has to form a bubble which is not that easy by itself. A nucleation site makes it easier for the bubble to form; so more nucleation sites, more bubbles. Once a tiny bubble forms, it is easier for more CO2 to join this bubble. It is easier for more CO2 to join a bigger bubble than a smaller bubble.
The original container has nucleation sites but these have been "conditioned" by prolonged exposure to the carbonated solution. Nucleation sites in a different container also hold on to a bit of atmospheric air when the soft drink is poured in (aside from residual dust or other extraneous material providing additional sites) so these nucleation sites are more active. Also, smaller receiving containers have more sites per volume than larger containers. So there is more fizz in the smaller receiving containers and they go flat faster. The fizz, or rising bubbles, also stir the contents a bit. If we pour into a container with ice, there is thermal shock, more convection currents, greater exposure of liquid to air, and (believe it or not) more nucleation sites (quality of the ice water is probably a factor as well, hmmmm? i.e. more contaminants, more imperfections = more nucleation sites); most of these factors have been raised already.
If we can make the bubbles smaller, then the rate at which CO2 comes out of solution will be slower. Gas volume is inversely related to pressure. So pumping up the orginal container with air (even though it is not CO2), can slow down the loss of dissolved CO2.
That's how I see it.
Although this is a bit of fun, there are serious industrial and medical applications/problems related this subject.
Regards
Would you mind addressing the issue of pouring cola onto ice cubes (variation of amount of fizz with the size / shape of ice and other materials) and mixing milk with root beer? 
These issues were touched upon here, but never to the depth. I predicted that smaller ice cubes should create more fizz. What do you think?
