The Semantics of Derived Lexemes
When somebody makes up a word, they are inventing it for use under a particular circumstance. Sometimes the circumstance can be very peculiar. Take the sentence in (6):
(6) Joe was Houdini’d and died.
In order to understand this sentence, you have to know something about Houdini – a famous escape artist – and how he died. He died following a series of punches to his stomach (a fan was testing the strength of his abdominal muscles, which Houdini prided himself on). So when we say “Joe was Houdini’d and died,” we mean he was punched in the stomach and died in the way that Houdini did. This is a dramatic example of how you might need to know pragmatic factors in order to understand a particular lexeme. We say more about examples like (6) below.
The second factor that can affect a word’s meaning is its history. We might think of every lexeme not just as a word and its meaning, but as the word and every time it has ever been used: every time we hear the word, we revise its lexical entry in some way. That this indeed goes on is particularly evident from first-language acquisition research. Children in earlier stages of language acquisition may underextend a word by using it to refer to only a subset of its actual referents, or overextend a word by using it to refer to objects or individuals that are typically covered by the word, as well as to others that are “perceptually similar” (Clark 1993: 33). For example, a child might underextend the word dog by using it to refer to more typical examples of the species, but not to varieties like Chihuahua or Pekingese (Kay and Anglin 1982), or overextend tree by using it to refer to potted plants, trees, and even balsam fir wreaths. Such under- and overextensions are generally short-lived, which indicates that children revise lexical entries as they are exposed to more and more tokens of a word.
It is not unreasonable to think that the meaning of a word is a compilation of every single use of that word that you have ever heard or said. Every word has a history. It has your own personal history – how you have heard the word. It has the history of the word as it has been used by other people. Over time, the meanings of words can become more complex and diverse, making the task of the morphologist looking for semantic patterns of word formation more complicated than it would be if the semantics of word formation were purely compositional (as the semantics of syntactic constructions are often considered to be). A syntactic construction may have pragmatics to deal with, but it doesn’t have history.
One question you might want to ask is what kinds of meanings arise via lexeme-formation rules. Are derived forms like lexemes, with potentially very complicated meanings? Or are they like syntactic collocations, with simple meanings?
