the worst problem for us with the idea of giving the dictionary the authority to determine whether xyz is a word is that dictionaries don’t need to include every word. Every language has ways of forming new words that are so active and transparent that putting all the words formed that way into the dictionary would be a waste of space. For example, speakers of English know that any verb at all can have a present progressive form made with the suffix -ing. As soon as I make up a new verb, say zax, we know that the present progressive verb form is zaxing. So although a dictionary might eventually have to include the verb zax, it might never list zaxing as a word. But of course zaxing should be considered a word. Similarly, just about any adjective in English can be made into a noun by adding the suffix -ness. For example, the Concise Oxford English Dictionary contains the adjective bovine, but not the noun bovineness. Nevertheless, I’d have no problem if I saw the word bovineness written somewhere, and would never think to look it up in the dictionary. The dictionary doesn’t have the word precisely because we’d never need to look it up.

The conclusion that we are inexorably led to is that we cannot rely on dictionaries to answer the question “Is xyz a word?” On the one hand, dictionaries don’t list all the words of any language. They can’t list all derivatives with living prefixes and suffixes, or all technical, scientific, regional, or slang words. And on the other hand, they sometimes include words used only once whose meanings are completely unknown. They occasionally even include purposely made-up words to guard their own copyrights. For the most part, dictionaries do not fix or codify the words of a language, but rather reflect the words that native speakers use. Those words are encoded in what we will call the mental lexicon, the sum total of word knowledge that native speakers carry around in their heads. So to answer our question, we must look more closely at what is in that mental lexicon