I think this can best be grasped with a little history, if you can bear with me a moment.
Baud rate, or symbol rate, is the speed at which the basic data bit is transferred. It comes from Baudot code, which was the original 5 bit mark and space teletype code. With 5 bits of information, you have 32 characters available, which back ini the dark ages was enough for the upper case English alphabet, plus 6 control characters.
Standard "baud" rate for landline teletype was 45 baud, which translated to about 60 words per minute. The information rate is 60 words per minute, or about a word per second.
ASCII code, which uses 7 bits, gives you a lot more characters...128 in all...all the upper and lower case letters, all kinds of punctuation, and a few extra slots for things like Cyrillic characters or the like.
However, for a given "baud" rate, the character (or information) rate for ASCII is slower than for Baudot code, since each character now had to have 7 bits of information instead of five.
To get the one word-per-second rate using ASCII, you have to bump up the data or SYMBOL rate to about 100 baud. (ASCII also has a start bit which is not used in Baudot teletype)
So, you see, the more complex a code is, the slower the information rate is for a given baud rate.
Each binary bit is information. You don't have to have a whole character/word/sentence/etc to have information.
Baud is a count of the binary reversals of the signal, per second.
In a simple system, each baud carries one binary bit of information. Information rate = baud.
When you are able to transmit more than one binary bit per each reversal of the signal, you can get information rates higher than the baud rate.
For example: if each frame is divided into 8 or 16 or more divisions, and the position of the signal reversal within this frame can be measured, multiple bits of information can be transmitted with each reversal. Thus information rate > baud rate.
Ever wonder how phone modems can transmit 56K bits per second over a phone line that is frequency limited at about 3KHz? (or at least significantly less that 56Khz). It does it by transmitting many multiple bits per baud. Information rate > baud rate.
There are many other ways to get information rate > baud rate.
Baud rate (modulation) rate is the rate at which a signal is modulated is the rate at which the symbol change. The modulation rate is known as the baud rate and has the units of baud. The modulation rate is the reciprocal of the symbol time.
Information Rate (bit rate) is the number of bit information transfer per second.
Relationship :
Bit rate = Baud rate χ (Bits/symbol)
= (symbols/seconds) x ( bits/symbol)
= M log2L.
Also don't forget that standard serial 8N1 has an extra start and stop bit per 8 data bits, so the ratio is 10:1 so it takes 10 "bauds" to make 1 byte.
And CONGRATS on looking in a book. Books are good.