Every sentence spoken in every language is built from grammatical atoms called morphemes. A morpheme is the "smallest meaningful unit" you can find when you break phrases and words apart. Morphemes can be long like annihilate or very short like the "past tense" -(e)d tacked onto annihilated.
Morphemes have allomorphs, or various ways they show up in language. In English, -ed marks the past tense, but it doesn't always have the same pronunciation. Speakers pronounce it as [t] in helped but [d] in cubed. Past-tense [t] and [d] are simply two allomorphs of the same morpheme /d/.
You now have some basic concept of morphemes and allomorphs, but how do you apply this to everyday language? To begin analyzing words, separate morphemes with a hyphen.
| English | Amazing! The farmer speaks Latin! |
| amaz-ing the farm-er speak-s Latin |
Keep in mind that morphemes and allomorphs are identified based on how speakers of a given language build words. Morphemes are language-specific. English speakers cannot break annihilate into smaller meaningful units, but Latin speakers could find the historical morphemes an-nihil-at-e.
Types of morphemes
Different morphemes have different functions. The morphemes farm, speak and amaz(e) have clear semantic content - if you speak standard English, you know what they mean. But what about less obvious morphemes like -s and -ing? These morphemes attach to another morpheme to provide it with a grammatical meaning, and are known as inflectional morphemes. The grammatical meaning of the inflectional morpheme -ing in amaz-ing is “present participle”.
The case of -er in farm-er doesn't involve the kind of "grammatical meaning" expected from inflectional morphemes. Instead,-er creates a new word with a new meaning. While speaks and speak may be considered a separate form of the same word, farmer and farm are two different words. Morphemes like -er that attach to other morphemes to derive new words are known as derivational morphemes.
Notice that morphemes like speak and Latin can exist on their own - they are free morphemes. On the other hand, bound morphemes must attach to another morpheme before a speaker can use them. Inflectional and derivational morphemes cannot exist on their own - they are known as affixes (from the Latin stem af-fix 'fixed to'), and are necessarily bound morphemes.
Affixes that attach before other morphemes are called prefixes, including the derivational pre– in pre-fix. Suffixes attach to the end of another morpheme, such as the inflectional -ed in studi-ed. Since they fall at the end of words, suffixes are also referred to by the informal name endings. When written alone, you may see prefixes with a hyphen following the morpheme and suffixes with a hyphen preceding the morpheme.
Base morphemes and attaching morphemes
The morpheme that carries the word's central meaning is often called the root morpheme. In French, the root port- has the meaning 'carry'. French speakers must add inflectional affixes to that root to make it usable, such as porter 'to carry', portant'carrying' or elle porte 'she carries'.
French speakers may add derivational prefixes to the root port- to form ap-port- 'carry towards'/'bring' and its opposite em-port- 'carry away'/'take'. These are derived words, but we cannot call them roots - they are stems formed from the root port-. Again, French speakers must add inflectional affixes to that stem to use it, as in apporter 'to bring' or elle apporte 'she brings'. The morpheme port- is the root and stem of porter, while apporter has the root port- and the stem apport-.
Ways of expressing grammatical concepts
In this course and especially in the parallel coursebook Native Grammar, I will give examples from a variety of languages. Each language handles grammar differently, but we can remark on a few general cross-linguistic tendencies.
A language may grammatically mark a word by attaching an affix or placing it near another word with some grammatical meaning. The French word porté 'carried' is grammatically marked, but so is the Japanese word 日本語 nihongo 'Japanese (language)' in 日本語の本 nihongo no hon (Japanese - possessive affix - book) '(a) Japanese book'. Words that do not take such morphemes are unmarked.
Keep in mind that unmarked or less marked words may also represent a grammatical concept: the word mom is an unmarked singular noun (the plural mom-s is marked). Since in this case nothing means something, we can talk about azero morpheme or null morpheme, sometimes written as a zero (0). Applying this concept, mom-0 represents the singular (only one mom) while mom-s represents the plural (two or more moms).
A language might explain a feature in words rather than mark it outright. The marked Romanian word omului has the rootom 'person' and an affix –ului that marks direction towards. English uses periphrasis, Greek for 'speaking around', to translate this single Romanian word: 'to the person'.
A language may also rely on compounding, which involves combining two or more words to form a new concept:keyphrase simply compounds the noun key with another N phrase.
A language may simply appropriate a word from one word class and use it in another, a technique known as conversion. For example, the noun query (class = N) has also come to be used as a verb query (class = V) without undergoing any change in form: query the database.
Lastly, a language may use word order, or arranging words in a specific way. This is an especially important when considering basic word order, which describes the normal arrangement of subjects, verbs and objects in a language. So, English may be considered an SVO (subject-verb-object) language, Classical Latin an SOV language, and Classical Arabic and Modern Irish a VSO language.
Classifying languages by their word-building tendencies
Some languages, including Indo-European languages like Latin, Greek, and Sanskrit, tend to form words by tacking on all kinds of grammatical affixes. We describe these languages as synthetic languages. Synthetic languages build words by adding affixes to roots and stems. Synthetic languages are associated with bound morphemes, especially when it comes to content words like nouns and verbs. Consider Caesar’s famous quote, in which the verbs ‘come’, ‘see’ and ‘conquer’ are necessarily bound to an affix:
| Latin | Ven-i vid-i vic-i. |
| 'I came, I saw, I conquered.' | |
| Ven-io vid-eo vinc-o. | |
| 'I come, I see, I conquer.' |
There are two major types of synthesis. Fusional languages shove a lot of grammatical information into their affixes. The ending -i in the Latin examples above fuses together information about who conquers and when the conquest takes place.
More commonly, synthetic languages divide grammatical information into separate affixes. Such agglutinative languagesoften add strings of affixes to form words.
| Turkish | Ev-iniz-de-yiz. |
| (house-your-at-we) | |
| 'We are at your house.' |
Many languages are not synthetic at all, and require few if any affixes. Such languages are analytic. In such languages, the vast majority of morphemes are individual, usable words in their own right. Modern English has this tendency: learn is a root morpheme and a word in its own right, while equivalent Italian root impar- must take on affixes that turn it into a usable word, such as imparo 'I learn' or imparare 'to learn'. Compare this to a Mandarin Chinese word like 喝 (hē) 'drink', which never takes affixes. Analytic languages may employ periphrasis, compounding and word order to convey the kind of information explicitly marked in synthetic languages.
Crucially, all of the terms in this section represent general tendencies rather than exceptionless rules. Languages aren't purely synthetic or analytic. Synthetic languages have words that cannot attach to affixes, while analytic languages may use separate words to mark grammatical function in a way strongly reminiscent of affixation.
Practice Exercise
1) Divide the following phrase into morphemes.
I reread the book's ending but it still doesn't make sense.
2) Which morphemes are free and which are bound in the sentence above? Identify the roots, stems and affixes.
3) Based on that same sentence, how would you classify English (among the types of languages introduced above)?
4) English speakers may talk about many books (with an [s] sound) but many songs (with a [z] sound). What is the relationship between the suffix pronounced [s] and the suffix pronounced [z]?
