Phonemes—the topic of my last two posts—are the smallest bits of language that make a difference. A phoneme doesn’t mean anything, it just allows us to tell one word (or morpheme) from another. So the phonemes /p/ and /b/ allow us to tell the word “pat” from the word “bat”. Phonemes are more or less the size of a letter in an alphabet. More or less, but not exactly.
Morphemes are more or less the size of a word, but not exactly. Many words in English are morphemes, but some morphemes are smaller than a word, and some are bigger (and some languages don’t have words the way English does, anyway). Some morphemes can exist as independent words, but some have to be attached to other morphemes. The general study of morphemes is called morphology.
A morpheme is the smallest bit of language that means something. There are zillions of morphemes in English. “Dog”, “cat”, “unicorn”, “cabbage”, “university”, “soccer”, “golf”, “gold”, “child”, “eat”, “sing”, “brown”, “towards”, “the”, “very”, “which”, “why”, and so on and so on and so on. An English dictionary is mostly a gigantic list of word-sized morphemes. We can categorize these word-sized morphemes into classes that we call the parts of speech: in English the parts of speech are nouns, verbs, adjectives, adverbs, articles, prepositions, conjunctions, and exclamations. In earlier posts we have seen that these categories are kind of fuzzy, but they will do for now.
The word “dog” is a morpheme. It’s a small bit of language and it has a meaning and it can’t be carved up into smaller bits that have meaning. A simple sentence such as “I love my dog” has one meaning, and the sentence “I love my cat” has a different meaning, and it’s the difference between the morpheme “dog” and the morpheme “cat” that makes the difference. (For now we will ignore some possible complications, such as metaphorical meanings and uses of the word—“My dogs are killing me”—or the use of the word as a verb—“I will dog your steps”—but any good account of meaning will have to return to such meanings at some point.)
If the word “dog” is a morpheme, what about the word “dogs”? This is made of two morphemes—the morpheme “dog” and an added morpheme “s”, which makes the word plural. We can say that the “s” here is a plural morpheme—it has a meaning, which is “plurality”. But this “s” is not the only plural morpheme. The “s” in cats is a different plural morpheme; it may look the same, but in the word “cats”, the “s” is pronounced “s”, while in the word “dogs”, the “s” is pronounced “z”. And in the word “horses” the plural morpheme is pronounced “ez” (more or less). These are three allomorphs of the plural morpheme, in somewhat the same way that aspirated “p” and unaspirated “p” where allophones of the phoneme /p/. As a general rule, the “ez” plural occurs after the sounds “s” “z” “sh” “zh” “ch” and “tz” (dresses, hoses, flashes, rouges, hatches), the “s” plural occurs after the sounds “p” “t” “k” “f” and “th” (caps, taps, backs, gaffes, cloths), and the “z” plural occurs everywhere else. This is another one of those rules that everyone follows without knowing about.
In addition, the English plural morpheme has a few other allomorphs, though these are less productive, such as the contrast “man” and “men”, “woman” and “women”, “foot” and “feet”, “child” and “children”, “ox” and “oxen”, “goose” and “geese”, or “sheep” and “sheep”. This last example shows what is called the zero allomorph, since the singular form and the plural form are the same. Morphemes are conventionally enclosed in curly brackets; if we write the general sense of the morpheme, we can write in capital letters, {PLURAL}, or we could write out the allomorphs {s, z, ez, etc.}. The pronunciation of the stem can change as well: we pronounce “child” as /chaild/ (more or less) with a diphthong, but “children” is /child-ren/. We can say that the morpheme {CHILD} has the allomorphs {/chaild/, /child-/}. Many morphemes, both roots and affixes, have allomorphs. For instance, the root “knife” becomes “kniv-” in the plural “knives” and “leaf” becomes “leav-” in the plural “leaves” (except in the team name Toronto Maple Leafs). The negative prefix “in” (“incredible”) can show up as “im” (“impossible”) or “il” (“illegal”) or “ir” (“irregular”) or “ing” (“ingconsistent”—if you listen carefully, you can hear that “in-” before words that start with “c” is usually pronounced “ing”). The prefix “syn” (“synonym”) can show up as “sym” (“symphony”).
Two roots can be compounded to form what looks like a single word, as in “screwdriver” or “blackbird” or “blueberry”. It is not always easy to tell if a compound is one word or two, and spelling may be somewhat arbitrary: “baseball” is one word, but “base hit” is two. Some languages form compounds easily; German is famous for its long compounds, which are easily invented on the spur of the moment. German dictionaries don’t even try to keep up, so looking up a word in German often means looking up its various parts and deducing what the compound must mean.
Morphemes in a sense are all about meaning. But even if a morpheme is the smallest bit of language that means something, that doesn’t mean that the meaning of a morpheme is the smallest bit of meaning. The English word “uncle” is a morpheme, and you can’t divide it into parts, the way you can divide “dogs” into a root “dog” and a suffix “-s”. But the meaning of the word seems to be complex: “The brother of one’s mother or father”. This complex is, I suppose, a unitary meaning in English, and that’s why it’s bundled up in a single word. Some languages, however, divide this concept up into two different morphemes: in Latin your mother’s brother is your “avunculus” and your father’s brother is your “patruus”. In English we have a single word “river” where French has two: “fleuve”, a large stream that flows into a sea or the ocean, and “rivière”, which flows into a fleuve. So the Missouri River is a rivière and the Mississippi River is a fleuve. But English makes distinctions which other languages don’t make; for instance, English distinguishes “sheep”, the living animal, from “mutton”, the same animal served up for dinner, but French uses “mouton” for both. English has two words, “experience” and “experiment”, where French has one, “expérience”.
Some linguists have attempted to carve meanings up into fundamental components, semantic features, which are present, not present, or not applicable. Thus “man” is +male and +adult and +human, while “woman” is –male and +adult and +human, and “boy” is +male and –adult and +human, while “girl” is –male and –adult and +human. (You can see some of the trouble we’re headed for: why is “male” the default category? Does every selection of a default category hide some ideological presupposition? Can the world always be reduced to binary distinctions?) This method, which is called “componential analysis”, is derived from analysis of phonemes into component features. In my previous post I hinted at a sort of first step in the componential analysis of a few phonemes: the phoneme “p” is an unvoiced bilabial stop, while “b” is a voiced bilabial stop. Feature analysis of phonemes works quite well, but componential analysis of meanings is full of problems and it has been largely superseded by more subtle methods of semantic analysis, such as prototype theory, which I will discuss in another post.
Some literary critics have used a kind of componential analysis in the description of characters. The first sentence of Jane Austen’s Emma seems almost designed as the basis for componential analysis: “Emma Woodhouse, handsome, clever and rich, with a comfortable home and happy disposition seemed to unite some of the best blessings of existence and had lived nearly twenty-one years in the world with very little to distress or vex her.” Emma has three plus characteristics: she is (+handsome), (+clever), (+rich). The other marriageable women in the novel can be rated on the same grid: Jane Fairfax is (+handsome), (+clever), (–rich); Harriet Smith is (+handsome), (–clever), (–rich); and Mrs. Elton is (+handsome), (–clever), (+rich). This kind of componential character analysis may have some use, but it is clearly just a first step. People are a lot more complicated than phonemes.
The Art of Reading Slowly 