New words are assembled from parts you already know
The basic recipe is recombination. English lets speakers attach familiar endings to familiar stems, so a word like "unTrumpian" is instantly understandable even if you have never seen it before [3]. The same logic produces coinages like "wordify" or "grantology"—new strings of letters whose meaning is predictable from their ingredients [1]. This is why novel words can exist at all: they are not invented from nothing but assembled from material the language already supplies.
Not every assembly is equally welcome, though. Some endings are "productive"—used very often to make new words—while others are rarely pressed into service that way. The suffix -er, as in "dreamer," is a workhorse for coining new terms, which is why a made-up word like "berater" feels more natural than "penciler" or "plantist" [1]. Productivity, in other words, is a measure of how readily a word-part lends itself to fresh combinations.
Why some new words stick and others fade
A novel word becomes part of your vocabulary through exposure and competition. When you learn a new word, it does not simply sit in a separate box; it starts competing with similar-sounding words you already know. Experiments with made-up words that resemble real ones (for example, a novel item modeled on "cathedral") suggest that this competition is part of how a new word gets integrated into the mental lexicon—the internal store of words you know [2]. Integration appears to depend on time and on how much exposure the new item gets [2].
Real-world usage shows the same pattern at scale. In an English newspaper, novel words appeared at a non-negligible rate, and many of them were used only once—so-called hapax forms [4]. That tells us something important: the existence of a novel word is often fleeting. A coinage can be perfectly understandable and still never become a lasting part of the language, because most one-off creations are not picked up and reused.
The brain treats productive new words differently
Brain recordings add a further twist. When people read novel words, the most productive kind—a new verb plus -er, like "tributer"—was processed significantly differently from other novel forms such as new noun-plus-er or noun-plus-ist words. This distinction showed up across temporal-lobe and prefrontal regions [1]. Existing, familiar words, by contrast, did not differ among themselves in the same way, which points to productivity as a property that matters especially when a word is new [1].
One caution about how to read that result: the study found that novel words and their existing counterparts did not differ significantly in processing overall [1]. So the brain is not treating new words as a wholly separate category. The more accurate picture is that novel words are handled much like known words, with productivity shaping the fine details of how a new form is understood.
So do novel words really exist?
Yes—but "exist" covers two different things. A novel word can exist as a one-time coinage that a speaker makes and a listener understands on the spot, and it can exist as an established entry that has been integrated into the mental lexicon through repeated exposure [2][4]. The first kind is common and often ephemeral; the second kind is what we usually mean by a word "existing" in a language.
The evidence supports both. Newspaper data show novel words being coined and appearing in print at a measurable rate [4], and laboratory work shows that novel items can be learned and integrated into the mental word store [2]. Meanwhile, the fact that speakers can produce and comprehend forms like "unTrumpian" without ever having learned them shows that the machinery for making new words is always running [3]. Novel words are not a marginal curiosity; they are what a living vocabulary is made of.
Sources used in this answer
Productivity matters for the neural processing of novel words, but not existing ones.
In MEG recordings, the most productive novel word form (novel verb-plus-er) was processed differently from other novel forms across temporal and prefrontal regions, while existing words did not differ among themselves, and novel words did not differ significantly from their existing counterparts.
Lexical competition and the acquisition of novel words
Experiments with novel words resembling real words suggest that integrating a new word into the mental lexicon involves lexical competition and depends on time and amount of exposure.
Introduction to English Morphology: words and their structure
English morphology allows novel words such as "unTrumpian" to be perfectly understandable because their meanings are predictable from their parts.
Chronicling the Times: Productive lexical innovations in an English newspaper
An English newspaper contained novel words at a non-negligible rate, including many that appeared only once.
Lexical creativity
Lexical creativity involves both very frequent and less frequent word-forming processes, and novel words arise on the basis of those processes.