1 and 2 and 3 and 4 and

Does the girl know computing addition to math? Probably.

"Here is an array of three character strings, indexed from [1]:

s[] = { "Yes", "No", "I don't know" }

what is the value of s[x] where x is the number you are thinking of?"

Basically, the space of three possible answers can be used as symbols to encode the information directly.

Justification, in the light of comments:

The other answers differ in that they employ an arithmetic and logical coding trick: arithmetic is applied and then logic to produce an answer, whose truth value or in determinacy is then rendered to English "Yes", "No" or "I don't know".

It is just as valid and "mathematical" to simply obtain these symbols directly without using arithmetic coding.

Furthermore, it can still be regarded as arithmetic coding, because the answer strings are made of bits, and can therefore be coded as numbers: for instance, the bit patterns of the ASCII characters can be catenated together and treated as large integers. s is then effectively just a numeric table lookup which maps the indices 1 through 3 to integer symbols which denote text when broken into 8 bit chunks and mapped to ASCII characters.

A lookup table, though arbitrarily chosen, is a mathematical object: a function.

Furthermore, the displacement calculation to do the array indexing is arithmetic; we are exploiting the fact that the information we are retrieving is numeric and can be used to index into a table. Otherwise we would have to specify an associative set relation instead of a function from the integer domain. ("Here is a mapping of your possible state values to the symbols I'd like you to use to send me the value.")

This answer reveals that the question is basically uninteresting. An entity holds some information that can be in one of three states, and there is to be a three-symbol protocol for querying that information. It boils down to, give me the symbol which corresponds to your state, according to this state->symbol mapping function. I would therefore argue that the convoluted arithmetic coding is the hack answer not this straightforward coding method. In computing, we sometimes resort to arithmetic encoding hacks when we have to use a language that isn't powerful enough to do some task directly, or simply when the resources (time, space) aren't there for the cleaner solution.

What is it called when you add 1 2 3 4?

For those of you who are unfamiliar with this series, which has come to be known as the Ramanujan Summation after a famous Indian mathematician named Srinivasa Ramanujan, it states that if you add all the natural numbers, that is 1, 2, 3, 4, and so on, all the way to infinity, you will find that it is equal to -1/12.

What is it called when you do 1 2 3 4 5?

Natural Numbers (N), (also called positive integers, counting numbers, or natural numbers); They are the numbers {1, 2, 3, 4, 5, …}

What is 1 * 2 * 3 * 4 * 5 all the way to 100?

➡ 5050. Sum of 1+2+3+4...... 100 is 5050.

Is the Ramanujan summation true?

It assumes that this sum has a well-defined value, on which standard operations (addition, subtraction, mulitplication, division) are then defined. But this is obviously untrue.