A morpheme is the smallest unit of language that carries meaning. A word consists of one or more morphemes. For example, elephant is a morpheme, the plural suffix –s is a morpheme, the word elephants is a word consisting of two morphemes.
Morphemes can be “free”, meaning they are used by themselves, (e.g. house, car, walk, eat, from, to, but, him) or they are “bound”, meaning they only appear when attached to another free morpheme (e.g. the plural suffix –s, the past tense –ed, the prefix re-, the possessive ‘s, etc.). The concept of a morpheme is extremely useful in describing any language, and I’ll make use of it throughout this post.
Traditional analysis of English has a heavy focus on words. The eight parts of speech, which is a core concept in traditional grammar, is all about classifying individual words. Many prescriptive rules are concerned with which words to avoid: Don’t end sentences with prepositions. Don’t start with conjunctions. Don’t say ain’t. Don’t use singular they. Nearly every “improve-your-grammar” book has a list of commonly confused or misused words: their/they’re/there, who’s/whose, that/which, principal/principle, etc.
However, languages are not just big bags of words. Words work together, and they can be grouped into larger units called “phrases”. Phrases are an essential part of any theory of grammar. It’s impossible to properly describe a language without at some point needing to talk about groups of words, instead of just individual words. The most obvious example of such a unit is the “sentence”, which is actually just a very large phrase.