A word and its parts: roots, affixes and their shapes

Taking words apart

There are many words that need not be listed in dictionaries, because their meanings are completely predictable (such as dioeciously), and many which cannot be listed, simply because they may never have been used (such as un-Clintonish and antirehabilitationist). These are all words which are not lexical items. But what is the basis of their semantic predictability? It must be that these unlisted and unlistable words are composed of identifiable smaller parts (at least two), put together in a systematic fashion so that the meaning of the whole word can be reliably determined. In un-Clintonish these smaller parts are clearly un-, Clinton and -ish; in dioeciously these parts include dioecious and -ly, with further smaller components being perhaps discernible within dioecious. We will focus on these smaller parts of words, generally called morphemes. (The area of grammar concerned with the structure of words and with relationships between words involving the morphemes that compose them is technically called morphology, from the Greek word morphe ‘form, shape’; and morphemes can be thought of as the minimal units of morphology.) We will be concerned with two important distinctions between different kinds of morpheme, and we will consider ways in which a morpheme can vary in shape.

Before we embark on those issues, however, there is an important point to be made concerning the distinction between words that are lexical items and words that are not. As we have seen, words that are not lexical items must be complex, in the sense that they are composed of two or more morphemes. But those are not the only words that are complex; lexical-item words can be complex too. To put it another way: words that are lexical items do not have to be mono-morphemic (consisting of just one morpheme). This is hardly surprising, when one considers that we have already encountered lexical items that are so complex as to extend over more than one word, namely idioms. But recognizing the existence of lexical items that are poly-morphemic (consisting of more than one morpheme) has an important bearing on the relationship between morphemes and meaning, as we shall see.

Let us look in more detail at two characteristics of morphemes, in the light of how the notion has been introduced. To allow the meanings of some complex words to be predictable, morphemes must

1. be identifiable from one word to another

and

2. contribute in some way to the meaning of the whole word.

Now, what permits the same morpheme to be identified in a variety of different words? A morpheme cannot, after all, be just any recurring word-part. To see this, consider the words attack, stack, tackle and taxi. These all contain a syllable pronounced like the word tack; but it would be absurd to say that the same morpheme -tack- is identifiable in each, because the meaning of tack has nothing to do with the meanings of the other words, and all of them must surely be listed separately in any dictionary. So it may seem natural to link characteristic 1. tightly to 2., making the identification of morphemes dependent on their meaning. Indeed, in introductory linguistics textbooks, one often encounters statements to the effect that morphemes are not merely the smallest units of grammatical structure but also the smallest meaningful units. This view is widespread precisely because it fits many complex words very well – not only brand new words like un-Clintonish but also established words like helpfulness, which is divisible into the morphemes help, -ful (identifiable also in cheerful and doleful, for example) and -ness (identifiable also in happiness and sadness). It seems reasonable to say that the meaning of both un-Clintonish and helpfulness is entirely determined by the meanings of the morphemes that they contain. Even the meaning of a word such as readable, which  is idiosyncratic enough to require mention in a dictionary, is clearly related to the normal meanings or functions of read and -able. In the face of such examples, it is important to remember that there is no necessary or logical connection between characteristics 1. and 2. Repeatedly, we will encounter evidence that it is risky to tie the identification of morphemes too closely to their meaning.

Another general point to be made about morphemes is that, although they are the parts out of which words are composed, they do not have to be of any particular length. Some relatively long words, such as catamaran and knickerbocker, may consist of just one morpheme; on the other hand, a single-syllable word, such as tenths, may contain as many as three morphemes (ten, -th, -s). What this shows is that the morphological structure of words is largely independent of their phonological structure (their division into sounds, syllables and rhythmic units). This reflects a striking difference between human speech and all animal communication systems: only speech (so far as we know) is analyzable in two parallel ways, into units that contribute to meaning (morphemes, words, phrases etc.) and units that are individually meaningless (sounds, syllables etc.). The implications of this property of human language (its so-called duality of patterning). What matters here is just that you should avoid a mistake that beginners sometimes make, that of confusing morphemes with phonological units such as syllables.