Why did people start using CO2 (instead of e.g. oxygen) for carbonated drinks?

CO₂ tastes good! Carbonated water is actually a weak carbonic acid solution; this is what you taste. It's like "salt and pepper", but for water.

Slightly acidic water is usually popular, hence lemon juice is sometimes used to taint a jug of water in times when no carbonated water is available.

Carbonated water releases CO₂ bubbles for many minutes after pressure is released (dissolved carbonic acid reforming), which is an important presentation effect that other gases do not generally have.

In many countries the local water supply is slightly alkaline (due to limestone rock being very common), and this can leave a bad after-taste, and make food in your mouth taste soapy. The addition of weak acids to water helps with neutralising this effect.

Around the world, some lucky towns have natural springs with effervescent water (carbonated water), and man-made carbonated water is just imitating this. In history people have often travelled great distances to try effervescent water, so its popularity is well recorded.

Nitrogen is used for water pressurisation in brewing and cooking as it leaves little or no taste, and generally does not form other chemical compounds. It's used mainly when bubbles are required, but not an acidic taste.

Oxygen is unsuitable for water pressurisation as very little can be dissolved at normal soda pressures, and it has no taste, but may readily form unwanted chemical compounds.

In the 18th century, Joseph Priestley was the first person to document this manufacturing discovery. He noted precisely that it tasted good, and that it was similar to the German Selters water. He was discovering "airs" (gases), and uniquely identified quite a few different gases. CO₂ was the only one he listed as tasting good. His cheap source of CO₂ was the fermenting grains from the nearby brewery.

Brewing processes often introduce carbon dioxide naturally, usually along with alcohol.

Carbonated beverages get the fizz and some spiciness/acidity without the alcohol.

Oxygen is MUCH less soluble in water than carbon dioxide. This is due to formation of carbonic acid. There would certainly be much less "fizz" if soda was charged with oxygen under the same conditions of pressure and temperature.

Also it might be undesirable to have high oxygen concentration in the sealed drink as this might reduce shelf life. In certain ways carbon dioxide can act as a preservative.

All answers posted so far sound good.

An interesting addition to the answers above: There's such thing as nitrogenation, where nitrogen is used along with CO2. This is done mostly to replicate the flavor of beer with less carbonation, which is common in places such as England where beer is served at 55F. CO2 becomes less able to dissolve with temperature increase, therefore higher temperatures mean less carbonation. Many beer drinkers say that carbonation makes flavors harsher, whereas nitrogenation replicates "traditional" beers where the carbonation affects the taste less. It also has effects on mouth feel and appearance.

A nitro beer will taste a bit flatter, but will make up for that with a full, creamy mouthfeel.

source: Beer On Nitro - A Brewer's Explanation

CO2 isn't the only thing used to add effervescence to beverages. It's just the most common.

To be a little more historical (why people started using CO2), I'll say that bubbly beverages all originated from fermentation. Even things that we have as non-alcoholic beverages - root beer, ginger beer, used to be fermented a bit (and still are, by people that make their own.)

The flavors that we have come to enjoy were developed in the context carbonic acid, especially the sweetness. And as others have pointed out, oxygen is bad for flavors. Beer in particular is affected by oxygen, the yeast will generate sour/cidery flavors in the presence of oxygen and other stale flavors appear.

Side note: artificial addition of nitrogen to beer is a recent thing meant to mimic traditional cask ale, where carbonation is low and air is forced into the cask to push the beer out. In the short time before the oxygen damages the beer, the nitrogen will dissolve in the beer giving smaller bubbles and creamier foam.

Man tends to follow nature.

Water passing through limestone ( CaCO3 ) dissolves a tiny amount of the limestone and flavors the water.

If that water is artesian, it can have more dissolved Limestone, but as the water exits the ground, into a condition of lower pressure, the dissolved CO2 is released ( forms bubbles ).

None of these answers is actually quite right (though some come close). First, it has nothing to do with simulating natural fermentation products or naturally occurring liquids. Nor does it have anything to do with safety in transport. Nor does it have much to do with taste (especially since, as others have already pointed out, the sensation of soda has very little to do with taste, and much more to do with stimulating nociceptors in the mouth). CO2 is used because it is one of the only non-toxic, easily and cheaply producible gases that can remain dissolved in significant concentrations in water. while about 1.75g of CO2 can be dissolved in a kilogram of water at room temperature, only about 0.02g of nitrogen can dissolve in the same amount of water. The situation is not significantly better for oxygen, where only about 0.04g per kilogram of water will stay in solution.

I would think the Oxygen would chemically interact with the sugars and other components of the soda, and in effect burn the contents!

Adding my 2 cents: Having constructed a coke bottling plant (not operated), I'm aware that tiny amounts of activate carbon is an ingredient and the reason given was that it would react with any dissolved oxygen thereby inhibiting bacterial growth in the sugared drink.

In fact, if you were to drink fresh coke from a bottling plant, it's a bit harsh on taste. This, I'm assuming here; is because the activated C in the drink initially forms concentrates of CO3 ions before it squeezes out the remaining O ions from the solvent to stabilize to dissolved and evenly spread CO2. I must admit, that I couldn't balance the equation here.

But I do know that if you mix coke with say a bacterial fermenting item like say yogurt or beer; cap it and rock it; it could result in a small and sudden explosion. The trick to bottling coke lays in carbonating it before any bacterial formation could take place.