It seems not yet fully recognized that the tendency to non-repetition is a linguistic universal; that is, it is part of the basic structure of Language, independent of the type of language or dialect, of the historical period and of the stylistic level, and it is found in phonetics, morphology (including word formation) as well as syntax.
It is a characteristic feature of this general tendency that it always exercises its influence, and when we find cases where this influence seems to show no effect, the reason is that some other linguistic function, of a more specific type, has overthrown the non-repetition.
In order to demonstrate this kind of universality, we should really examine all the world's languages, from the semantic level to the phonetic. Until now our investigation has covered about 40 languages, mostly text material, but also dictionaries and word lists.
When you say a word twice or more, it is commonly called repetition, the term indicating that the same word has been said twice, in other words, that the same event occurred two consecutive times. It is not difficult to understand that this is not possible (and already Heraclitus did understand it and probably some people before him): only if you take the event out of its context, consisting of time and place, etc., you can talk about repetition. Otherwise, one must take into consideration, in such a case, that a word is said at a time and in a particular place to a particular person, along with other words, etc., and therefore, by definition, it is not possible to make the same identical action again unless you turn back the clock and that has not yet succeeded. But in addition there is another factor or rather, a particular aspect of the problem mentioned above: the fact that an action is number two, e.g. that a word appears once, and then again, does exactly that it is not the same word the second time: the second occurrence is influenced by being number two, that the "same word" has already been said once before; occurrence no. 2 refers to the occurrence no. 1 (through the formal identity), whereby both no. 1 and no. 2 become something different than one and the same word, namely no. 1 and no. 2: two words that apparently are one and the same when you do not pay attention to the fact that no. 1 is alone (so far that is until no. 2 appears) and does not refer to anything before it (in this context), while no. 2 on its arrival already is known, you recognize it and do not regard it as a new element (as you did with no. 1), but it is automatically perceived, defined as a copy of no. 1 (which is known in advance) that comes back, so that no. 2 is first and foremost something different from itself, namely no. 1, that has come back, and is looking for a new place, a new identity. In this way occurrence number two often contains a different meaning than number one, either it emphasizes one side of no. 1, or it restricts or expands no. 1, or in still another way it changes its content by its relation to no. 1.
In many cases, we need to mention the same thing or concept several times in a text. It is characteristic of technical prose that you refer to one or more topics many times over a relatively short text, simply because these items are the themes about which you say something, that's the kind of sentences the text consists of. Essays and Art prose on the other hand, favour higher features as beauty or elegance, which requires that massive repetition be avoided where they are not used in a stylistic way. On the contrary repetition is said to enhance the beauty and efficiency of texts with more artistic character as essays with general human science subjects or short stories and novels. The strongest use of repetition seems to be found in poems with rhyme, alliteration, assonance and the like, where the sound effects are prominent and characterize the text so much, that you can not even imagine the text without these distinctive phonetic structures. Here we are far away from the introductory talk about repetition that does not occur in the true sense. In these cases, the authors use the phonetic substance of words and expressions without complicating the matter by asking if repetition does exist. In a fairly technical text, the reader can thus accept that a word or phrase is repeated a number of times since repetition is justified ("The Graces represent .. the historical conditions of the Graces .."). If the author then at a time uses a homonymous word with a totally different meaning, it is no longer a necessary repetition, that the reader accepts because it seems justified, but a (at least apparently) fortuitous coincidence where repetition is purely phonetic and not justified by the need throughout the text to maintain a given concept. This unjustified accidental and (only) phonetic repetition that breaks the line of justified and necessary repetitions ("The Graces represent .. the historical conditions of the Graces ..; in this case the sculpture has no grace"), can easily disturb the reading to the same extent that do the not uncommon convulsive attempts to avoid justified repetitions, such as Atlee, .. the Chairman .. the Prime Minister, etc, where the reader may be in doubt about the identity of the concept.