Originally posted by: nineball9
There was a show on PBS (possibly Nova) a few years back which explored the concept of counting in animals other than humans. In one experiment, researchers in a protected park (in Madagascar, IIRC) used wild lemurs as their subjects. The researchers were familiar with the local lemurs and were able to individually identify them at a distance so the experiment was not repeatedly conducted using the same lemurs.
The experiments employed techniques similar to those used with human infants; investigating and measuring stimulus, interest and attention span, and habituation and boredom. Basically, the experiment involved a box with differing numbers of apples (or some other fruit lemurs eat). A researcher put an apple in the box (about the size of a cooler on its side) and allowed a wild lemur to see the contents from a distance. Next, the researcher closed the box cover so the lemur could no longer see the apple, and from behind the box and not seen by the lemur, the researcher added (or removed) another apple. The researcher then reopened the front of the box and observed the lemur's reaction. 
Measuring the lemur's attention span to varying stimuli - number of apples in this case - allowed the researchers to correlate the lemur's interest with the varying number of apples in the box. When a lemur was shown an apple, the lemur showed interest (food!) If the researchers closed the box, then reopened it with a single apple again, the lemur's interest declined; nothing had changed. On the other hand, if the researcher replaced the single apple with two apples, the lemur showed increased interest; the lemur recognized the number of apples was different. If shown two apples again, the lemur showed less interest; the number of apples had not changed. This process worked up to three (4?) apples - any larger number of apples did not increase the lemur's interest level. The process also worked in reverse; when going from three apples to two, or two apples to one apple, the lemur showed increased interest as measured by increased attention spans. (Eventually, the lemur would get bored and return to the forest.) 
By repeating the experiment with numerous lemurs in the park, the researchers determined that lemurs could differentiate between 1, 2, or 3 apples. Basically, lemurs, or at least this species of lemur, could count to 3! 
Without getting too verbose, numbers and numerals are symbolic concepts used by humans. For instance, the numeral "4" is just a bunch of lines and has no direct reference to four objects. However, those who use Arabic numerals obviously understand the relationship. Yet the Chinese or Mayan symbols for four objects would be completely unknown symbols to people who use Arabic numerals (excepting those who can read Chinese or Mayan).  
On the other hand, counting is a cognitive trait that some animals, (including humans), do possess.
Chimpanzees, with larger brains than lemurs, can also count. Some researchers claim chimps can do some basic arithmetic as well. Try poking around 
www.pbs.org for some additional info. Here's a 
Scientific American Frontiers link about chimp counting. You might also poke around Google or pick up an intro Psychology text for more information.