Electronic tongue contains a group of sensors. Each sensor responds to a given analyte but with a different degree. If you use a second anlayte, the response pattern will be different. Based on the response patterns, you will be able to distinguish between analytes. Electronic tongue refers for liquid analysis; if you use the group of sensors for gas phase sensing, then it is known as electronic nose.

Thank very much for your response.

Quite often you will find the "tongues" and "noses" paired with "neural networks". The reason is that this type of sensors often responds in ways which are difficult to describe analytically.

What we have at a table is usually not a set of sensors which are specifically  sensitive to one particular substance, but to a set of sensors which are all to some level responding to many of the substances you are interested in and even in some you aren't interested in. So they need to be calibrated by: (1) putting the sensor array in a set of well defined environments; (2) measuring the signal coming out;  and (3) trying to figure out some function which would give good approximation for the actual gas concentrations we are interested in while eliminating the effects of ones we aren't. interested in.

This is a rather difficult task => hence neural networks.  

Which makes the noses/tongues name even more appropriate in a way. :-) 

If we take the liquid sensor definition we use for "tongue", we should put the tongue have reflected the basic sensory function. For example, sweet, bitter, sour, salty taste, these should be electronic tongue sense and can distinguish.