A puzzle: disentangling lexemes, word forms and lexical items

I have set out to distinguish and elucidate different senses of the word ‘word’, and to show how they apply in English. The outcome is something of a paradox. Words as basic units of syntactic organization (the building bricks out of which phrases and sentences are composed) do not coincide exactly with words as items listed in dictionaries. Indeed, there are mismatches in both directions, there are items that need listing but are not words in the grammatical sense, and there are words in the grammatical sense whose meaning and behavior are so reliably predictable that they do not need listing. There is yet a third sense of ‘word’, in that items that are words in the grammatical sense (lexemes) may have more than one form, depending on the syntactic context. Yet the items identified by the three criteria resemble each other sufficiently closely so that, in everyday non-technical talk about language, we do not even notice the discrepancies. Why should this be so? Is it so in all languages, or is English peculiar?

These are large questions. On the other hand, given that they arise so naturally out of issues addressed in an introductory text such as this, it is natural to expect that there should be some general consensus among linguistic scholars about how they should be answered. Yet there is no such consensus – something that, as a linguist, I am ashamed to admit. This reflects the meagreness of the research effort that has been devoted to morphology, the lexicon and lexical semantics over the last fifty years, by comparison with the huge intellectual resources devoted to syntax and phonology. So, for want of a consensus and of concerted research, the best that I can offer by way of a reply is speculation – albeit speculation informed by research in inflectional morphology.