How do you know the temperature was steady for 1800 years? the most effective global temperature readings didn't start until 1979.
As Firefly mentioned, there are ice core samples that can give us accurate readings. There are also many other methods of inferring paleoclimates. I just did this for my senior thesis, so I will give you a mouthful about sediment dating ;-)
As sediments deposit on land and in ocean basins, they build upon previous layers that have been deposited over time. This is called stratigraphy, an example of which you can see here:
Now, each one of these layers would represent a different time period. By taking a sediment core in an area, you can see these different layers in the sample and date them back to when they were deposited by which organisms are found in there. For example, in marine sediments you can look for organisms that used inorganic substances from the water column to precipitate their shells. One such type of organism, is called a foraminifera. They are tiny marine protists that precipitate their shells from the calcium carbonate and other substances found in the water. Here is what a fossil foraminifera looks like:
My study species is very charismatic, am I right? ;-) So, what we can do now that we have collected these fossils, is run them through a mass spectrometer. This will tell us the exact proportions of different elemental isotopes that were present in the shell, thus giving us an accurate record of the seawater chemistry at the time. Now, the important element we want to pay attention to here for inferring temperature is oxygen. Isotopes of O16 and O18 are found in seawater in specific proportions to each other, depending on the temperature of the atmosphere. This is because when we have large oceanic ice sheets, they are formed with the lighter oxygen isotope, leaving a larger amount of the heavier isotope in the water. So the proportions that we find of these give us a measurement of the ice cover of earth during that time, and a reading of the temperature once the proportions are entered into an equation.
So basically, taking fossil samples and breaking them down for their oxygen isotope proportions tells us what the temperature of the ocean and therefor atmosphere was at the time.
The fossils that I looked at ranged from about 21,000-31,000 years ago. And we can go much farther back than that. So yes, we do have an effective global temperature record going back very far in time.
Probably more than you wanted to know ;-P
I think the global warming is in only a period of time where the global temperature rises.
Well, this is where we should be correcting the use of the term. We really should not call this phenomenon 'global warming', but 'global climate change'. Yes, as an average, the overall temperature of the globe is rising. However, it is also decreasing in certain areas. Global climate change can be better characterized by more extreme conditions like with the seasons. For example, in certain areas summers will become much hotter, while in other areas winter will become harsher and colder. It all is resulting in an overall warming of the atmosphere, but it does not affect every part of the earth in the same manner.