Words are the nuts and bolts of language. All of us rely on a huge repertoire of words each time we communicate. We assemble long lists of words (dictionaries) and have frequent debates about what exactly a word means (or doesn’t mean) and who has the authority to decide about such issues.
Morphology is interested in the internal structure of words, much in the same way that phonology is interested in meaning-distinguishing speech sounds (phonemes). We can break down words into smaller units by analyzing their structure and identify systematic processes that allow speakers to add new words to the lexicon and indicate grammatical information such as tense and number.
An example illustrates the point. Think about what information is contained in the word girls. Is it possible to break this word down into smaller structural units?
girls = girl + –s
It seems that girls can be broken down into two parts, the first of which refers to something in the world (a young female human being) and the second indicating a grammatical category – in this case number – and specifying plural.
The same approach can easily be applied to other kinds of words.
kicked = kick + –ed
While girls is a noun kicked is a verb, yet the same rules apply. Kicked can be segmented into the first part that describes a kind of action (kick) and the second part that adds the information past tense (-ed). Tense is another grammatical category that can be encoded morphologically in English.
Think about what kinds of words take which endings for a moment. Only verbs (talked, laughed, pushed, loved) allow us to add information about tense, whereas only nouns (girls, boys, zebras, chairs) permit marking number.
Let’s compare this with the another kind of example. The word coolness consists of two parts, giving us the same kind of formula as in the previous two examples.
coolness = cool + ness
However, things look different when we analyze the segments. Cool can have a whole range of meanings, but most commonly it is an adjective that describes a person or thing. But what about -ness? It does not indicate number or tense – in fact it contains no information about any grammatical category whatsoever. -Ness also does not indicate a specific thing, action or state. So what is it good for? Look at these example sentences:
Mike is a cool guy
Coolness is a good trait to have
The -ness in words such as coolness, hipness, sadness or vagueness seems to mean “having the attribute X” and adding it to an adjective apparently changes that adjective into a noun. There are many more endings of this type that affect word class (for example, by transforming an adjective into a noun) and that may change a word’s meaning to different degrees.
teach – teacher
insane – insanity
happy – happily
A teacher (noun) is someone who teaches (verb), insanity (noun) is the state of being insane (adjective) and happily (adverb) is the way in which you do something you are happy (adjective) with or about. We can also extend or even reverse the meaning of a word by appending something like re- or un-.
fill – refill
introduce – reintroduce
happy – unhappy
fair – unfair
In linguistic terminology the minimal parts of words that we have analyzed above are called morphemes. Morphemes come in different varieties, depending on whether they are
- free or bound and
- inflectional or derivational
Free morphemes can stand by themselves (i.e. they are what we conventionally call words) and either tell us something about the world (free lexical morphemes) or play a role in grammar (free grammatical morphemes). Man, pizza, run and happy are instances of free lexical morphemes, while and, but, the and to are examples for free grammatical morphemes. It is important to note the difference between morphemes and phonemes: morphemes are the minimal meaning-bearing elements that a word consists of and are principally independent from sound. For example, the word zebra (ˈziːbrə) consists of six phones and two syllables, but it contains only a single morpheme. Ze– and -bra are not independent meaning-bearing components of the word zebra, making it monomorphemic. (Bra as a free morpheme does in fact mean something in English, but this meaning is entirely unrelated to the -bra in zebra.)
Not all morphemes can be used independently, however. Some need to be bound to a free morpheme. In English the information “plural number” is attached to a word that refers to some person, creature, concept or other nameable entity (in other words, to a noun) when encoded in a morpheme and cannot stand alone. Similarly the morpheme -er, used to describe “someone who performs a certain activity” (e.g. a dancer, a teacher or a baker) cannot stand on its own, but needs to be attached to a free morpheme (a verb in this case). Bound morphemes come in two varieties, derivational and inflectional, the core difference between the two being that the addition of derivational morphemes creates new words while the addition of inflectional words merely changes word form.
The signature quality of derivational morphemes is that they derive new words. In the following examples, derivational morphemes are added to produce new words which are derived from the parent word.
happy – happiness – unhappiness
frost – defrost – defroster
examine – examination – reexamination
In all cases the derived word means something different than the parent and the word class may change with each derivation. As demonstrated in the examples above, sometimes derivation will not cause the world class to change, but in such a case the meaning will usually be significantly different from that of the parent word, often expressing opposition or reversal.
probable – improbable
visible – invisible
tie – untie
create – recreate
Independently of whether or not word class changes and how significantly meaning is affected, derivation always creates (derives) new words from existing ones, while inflection is limited to changing word form.
Inflection (the process by which inflectional morphemes are attached to words) allows speakers to morphologically encode grammatical information. That may sound much more complicated than it really is – recall the example we started out with.
The word girls consists of two morphemes
- the free lexical morpheme girl that describes a young female human being and
- the bound inflectional morpheme -s that denotes plural number
Examples for the morphological encoding of other grammatical categories are tense (past tense -ed as in walked), aspect (progressive aspect as in walking), case (genitive case as in Mike’s car) and person (third person -s as in Mike drives a Toyota).
You are likely to notice that
- overall, English grammar has fairly few inflections and
- some inflectional endings can signify different things and more than one piece of grammatical information at once
The first point can easily be demonstrated by comparing English with German, which makes more use of inflection. Compare the following two pairs of sentences.
Der Mann sah den Hund
Den Hund sah der Mann
The man saw the dog
The dog saw the man
If you focus on the meaning of the two German sentences you’ll see that it does not change, even though we’ve changed the word order. The man is still the one who sees the dog, not the other way around. By contrast, the English expression changes its meaning from the first to the second sentence.
Why is this the case? In the German example the definite article is inflected for accusative case (den Hund), telling us who exactly did what to whom. This allows us to play around with the word order without changing the meaning of the sentence. English gives us no way of doing the same. We are forced to stick to a fixed word order due to a lack of case inflection (except for personal pronouns). Languages such as Latin that indicate a high degree of grammatical information via inflection (so-called synthetic languages) generally have a freer word order than analytic languages like English which have only reasonably very few inflections and rely on word order to signal syntactic relations (another popular example for a strongly analytic language is Chinese).
Linguists use the term affix to describe where exactly a bound morpheme is attached to a word. Prefixes are attached at the onset of a free morpheme, while suffixes are attached to the end. Infixes – affixes that occur in the middle of a word – are very rare in English, a well-known exception being expletive infixation. While in English suffixes can be either derivational or inflectional (teacher, slowly vs. apples, kicked), prefixes are always derivational (untie, recover, defrost).
Morphs, morphemes, allomorphs
When you look at certain inflectional endings that occur in English, you’ll notice that they are often but not always predictable. Here are a few examples for the plural morpheme.
one car – two cars; one rose – two roses…
one mouse – two mice
one man – two men
one ox – two oxen
one sheep – two sheep
A vowel change (also called an umlaut plural) instead of a suffix marks the plural in mice and men, in oxen the suffix we encounter is rather exotic (meaning this word is virtually the only one that takes the -en ending) and in the last example there is no visible plural marking at all.
The fact that plural number in English can be marked with several different inflectional suffixes (-s, -en), by vowel change or by no (visible) change at all points to a distinction you already know from phonology:
a concrete part of a word that cannot be divided into smaller parts
the meaning-distinguishing, abstract dimension of morphs, e.g. something like the pluralmorpheme
different realizations of the the same morpheme, e.g. -s, -en and nothing for the plural morpheme in dogs, oxen and fish_
When linguists talk about the allomorphs of the plural morpheme they are referring to variants of the same functional element which do not impact meaning in any way. A plural is still a plural, whether encoded by -s or something else.
Base, stem and root
Finally, in order to make the segmentation of words into smaller parts a little clearer, we differentiate between the base, the stem and the root of a word in morphological terms.
The stem is the base with all inflectional suffixes removed, whereas the root is what remains after all affixes have been taken off. When doing computational text analysis stemming (i.e. removing all inflectional endings) is frequently undertaken in order to avoid counting different word forms (e.g. house and houses) as separate words.
- free morphemes
- bound morphemes
- free morphemes
- morph – morpheme – allomorph
- base – stem – root
- synthetic language – analytic language