Redundancy (information theory)

In information theory, redundancy is a property of messages, consisting of having parts that are predictable from the rest of the message and that, therefore, in themselves do not provide new information or "repeat" part of the information.

In numerous applications, as well as in natural languages, redundancy is a widely used strategy to avoid misunderstandings or decoding errors. Descriptively, redundancy constitutes a strategic communication factor that consists of intensifying, underlining and repeating the information contained in the message so that the noise communication factor does not cause a fundamental loss of information. It also means 'excess'.

Pleonasm

Pleonasm is a type of redundancy consisting of the repetition of apparently unnecessary lexical elements in a sentence. Some examples of pleonasms could be:

a. "Go upstairs".

b. Get down..

c. "Get out".

d. "Enter inside".

e. Dead body..

f. "Freedom of books".

g. Wet water..

Although in many contexts the previous sentences are pleonastic or redundant, it should be noted that there are motivations to emphasize the meaning pleonastically. For example, although sentence (a) referring to real movement is that 'it always goes up', it must be kept in mind that there are expressions such as tension goes up, prices go up,... that do not imply real movement of any object (and something similar happens with (b)). As for (c), although it seems that it always "goes outward" There are metaphorical expressions like get ahead, come out gracefully, come out blushing that do not really imply movement. Perhaps the attempt to distinguish literal meanings from metaphorical ones is what has led some speakers to introduce pleonasms.

Now, there are pleonasms that, because they are too obvious, become barbarisms:

• Cold ice
• Round circle

Redundancy of natural languages

Claude Elwood Shannon was one of the first to be interested in the statistical redundancy of natural languages, particularly the English language. His research based on the conditional entropy of texts written in English, proved that natural languages contain a statistical redundancy close to 80%. That does not mean that in English texts we could eliminate 80% of the characters and still remain intelligible: experiments prove that if just over a quarter of the letters are randomly removed it generally becomes impossible to completely reconstruct the text from the remaining context. Most of the redundancy is undoubtedly due to the restrictions of the language's phonological system, which rules out certain combinations and causes only a few of all possible ones to give rise to well-formed syllables in English.