UserPreferences

ThTheProblem

What is reduplication?

You might have heard the phrase, “I know how to spell banana, but I don’t know when to stop.” It is a funny little saying, which speaks of our inability sometimes to know when to quit. But the phrase can help us understand what reduplication is also.

When writing “banana”, you write a “b”, then an “a”, then an “n”, and another “a”. After writing the first “na”, you know you need to repeat those letters again, but after writing so many n’s and a’s, it’s easy to get mixed up.

Let’s assume that “banana” is derived from some root word “bana” (like “speaking” is derived from the root “speak”). Now let us invent some rule for finding the “fruit-form” of a word, say: “repeat all letters from the last consonant until the end of the word.” In the case of “bana”, to get the fruit-form you would repeat the last “na” and end up with “banana”.

Change those letters into phonemes (in this case they are one and the same), and you have reduplication. Basically, we’re taking a root word and applying some rule to it which duplicates some subset of the phonemes in the word.

Reduplication is not terribly common in English (“banana” is not real example reduplication). In some languages however, reduplication plays a crucial role in the grammar. For instance, in the language of the Chunmash, a Native American Tribe, reduplication is used to pluralize words (Raimy 148). In English, a form of reduplication is used to emphasize adverbs like “very” and “really”. You might say, “This thesis is really really good.” This type of reduplication (reduplicating the entire word) is appropriately enough called “total reduplication.”

A form of total reduplication is also used in English to imply a certain, sometimes more romantic, meaning of a word in conversation. For instance, you might say to your friend, “Do you like her or do you like her, like her?”

Perhaps the most [well formed] use of reduplication in English, borrowed from the Yiddish, is the practice of repeating a word after the first consonant sound with “shm”. You might say, “Jason-Shmason” or “Thesis-Shmesis.” This is a more complicated form of reduplication since we are inserting phonemes that are not found in the root (“shm”) and is discussed later in this thesis. [WWW]Writing Tips

So that is the linguistic phenomena/problem of reduplication, but why should this stuff be of interest to computer scientists? Let’s look at one way of [formalizing] the reduplication problem.

Think of words as set of phonemes and a set of relations between those phonemes. One important relation used in constructing words is the “precedence relationship”. In the word cat, the “c” precedes the “a” and the “a” precedes the “t”. You might right that out with arrows, like:

This looks a bit like a linked list. What if you wanted to represent the total reduplication of “cat”? You might add a new link from the “t” to the “c”.