. . . But there are no Such Things as Words
Twas brillig and the slithy toves
Did gyre and gimble in the wabe;
All mimsy were the borogoves
And the mome raths outgrabe.
Did gyre and gimble in the wabe;
All mimsy were the borogoves
And the mome raths outgrabe.
Remember the Jabberwocky's song in Carroll's Through the Looking-Glass? Pretty meaningless, huh? Still, it sounds "English" rather than, say, French or German or Italian. That is why it is so amusing: it sounds like perfect English yet we cannot understand it.
Actually, we understand quite a bit about the poem even though we don't understand it as a whole. For example, what do we know about toves? Well, we know that there are more than one of them and that those mentioned here are slithy, whatever that is. We know that slithy describes the toves as either like a slithe or having slithes. We know that these slithy toves gyred and gimbled, and even though we don't know what these actions are we know they are actions and something about when they took place. How do we know all this, not knowing what any of the boldface parts of the words mean?
The reason we know so much about the meaningless words in Carroll's poem is that some of the words and parts of words are, in fact, English. The English words are small, less prominent words, barely more prominent in their pronunciation than the parts of words, the prefixes and suffixes. These small words and parts of words (in plain type above) are expressions that tell us nothing about the world, but only about the grammatical categories of the English language; let's call them morphemes, just to have a term for them.
The nonsense parts of the words in the Jabberwocky song above are all noun, verb, and adjective stems, the main parts of words without prefixes or suffixes. Stems are also the linguistic things that refer to the world we live in. Let's call whole words and stems lexemes; they are anything we use in speech that names or refers to things in the real world. Morphemes will then be the words or parts of words that mark the categories of grammar which are the crucial stuff of language. That is why they control whether we sense a spoken or written utterance is English or French or German.
No one has ever been able to define the word word despite gargantuan efforts to do so. The Linguistic Concept of Word: An Analytic Bibliography by Alphonse Juilland and Alexandra Roceric lists 118 PAGES of books and articles (unsuccessfully) attempting to define the word word over the past 3 millennia. Why can no one define word? Maybe because words simply do not exist; rather, the sentences we speak are composed of lexemes and morphemes and these two linguistic objects differ too much to be subsumed under one concept.
The differences between these two linguistic elements are dramatic. Here are the major ones.
|Comparison of Lexemes and Morphemes|
|Refer to real world||Refer to grammatical categories|
|May be derived||Cannot be derived|
|Always associated with sound||Often not associated with sound|
|Sounds always predetermined||Sounds often depend on stem|
|Are infinite in number||Belong to a small, closed class|
Let's examine these differences. Lexemes, remember, are basically noun, verb, and adjective stems or whole words: violin, play, small are lexemes. They refer to things, actions, states, and qualities in the real world. Suffixes like -ing, -s, -ed, prefixes like re-, ex-, un-, particles like the, not, so and prepositions like of, for, to are morphemes. They refer to grammatical categories like Progressive Aspect (-ing: am going), Plural Number (-s: cars), Past Tense (-ed: walked), Negation (un-, not: unwanted), Definite (the: the cat), Possession (of: of my friend).
Lexemes like violin, play, and small are subject to derivation, a process which generates new words from others: violin > violin-ist, play > play-er, small > small-ness. Morphemes do not undergo derivation even if they are not affixes which mark derivations themselves. We do not find words derived from the, that, of, for, it.
Lexemes always comprise a sound plus a meaning but morphemes may have either one but not the other. For example, if -er means "someone who does something", as in runner, player, driver, what sound has that meaning in (a) cook, (a) guide, (a) cut-up? With these verbs, all we have to do is use them in a noun position, e.g. the cook guided us through the kitchen, and we know that the speaker is using the noun, not the verb. (Leads one to wonder why we have affixes at all, doesn't it?)
On the other hand, morphemes some time show up with sound but not meaning. Take the adjectives that end on -ic, e.g. histor-ic, meaning "having the property of history" from history. There are lots of these derivations: climat-ic, theatr-ic, semant-ic. But look at this series: dram-at-ic, spasm-at-ic, enigm-at-ic. The meaning of these adjectives is the same as that of the "syntactic" class: "having the properties of X", where "X" is the meaning of the underlying word. However, they contain an extra morpheme, -at. Why is that? Well, for some reason known only to the ancient Greeks, who enjoyed the same derivation, before adding -ic you had to insert a suffix -at if the stem ended on an [m]. So, when English borrowed these words from Greece, the rule came with it. The suffix -at probably had no meaning in Greek but it certainly has none in English, where it is simply required of all words borrowed from Greek ending on [m] when the suffix -ic is applied. So not only do we find morpheme meanings without morphemes, we find morphemes without morphemic meaning.
The sounds of lexemes are always predetermined and remain (more or less) the same wherever they occur in a phrase. The sound of a morpheme on the other hand, especially prefixes and suffixes may vary wildly. The future tense in Tagalog (or Philippino) is formed by adding a prefix. For example, bili is the verb stem meaning "buy" and bi-bili means "will buy". So, if kuha means "get", "will get" should be bi-kuha, right? Wrong. The future of kuha is ku-kuha. Hey, what's going on here? The verb meaning "laugh" is tawa; "will laugh" is ta-tawa. Catch on? What do you think the future of sulat "write" is? You are right if you guessed su-sulat. Tagalog reduplicates the first syllable of the verb stem to make a prefix. In other words, the sound of the prefix is not predetermined, but is determined by the sound of the stem. The morpheme here is not a sound but a rule for creating sound. Only morphemes may be rules for retrieving sound, never lexemes.
|Tagalog Future Tense|
|bili "buy"||bi-bili "will buy"|
|kuha "get"||ku-kuha "will get"|
|punta "go"||pu-punta "will go"|
|sulat "write"||su-sulat "will write"|
|tawa "laugh"||ta-tawa "will laugh"|
An important difference between lexemes and morphemes is that the former, but not the latter, belong to an open, unlimited class. That means that there is no limit on the number of noun, verb, and adjective stems in any language. New lexemes may be derived by compounding, affixation, and other rules; they may be borrowed from other languages or simply made up. The number of morphemes, however, is limited to around 200. Proto-Indo-European languages have fewer than 100. We are currently being inundated by new lexemes describing the categories, states and qualities of new technology. We never meet a new preposition (of, for, by, with), pronoun (he, she, it), or article like a or the. The fact that the number of morphemes and the categories they refer to is fixed suggests that they determine the grammar of a language. That is why the Jabberwocky Song 'sounds' like English even though it is incomprehensible. It is English with nonsense lexemes but real morphemes.
So, speaking is a bit more subtle than stringing 'words' together in phrases. Assuming that on is a word does not distinguish between the lexeme on in the oven is on, and the morpheme on in on the table. Why is this important? It helps us understand the important difference between grammar and vocabulary. The meaning of on the morpheme, but not that of the lexeme may expressed by a suffix in some languages, such as Finnish and Hungarian. It helps us understand why children begin learning lexemes around the age of one, about a year before they begin learning grammar, represented by morphemes. They only begin picking up morphemes around the age of two, when they begin using phrases. Finally, it helps explain why chimpanzees can learn what seem to be lexemes but not morphemes (morphemes represent grammar, the essential core of language). Morphemes convey the categories of grammar itself, so if we want to understand language itself, we must understand first and foremost morphemes. Lexemes are not going to tell us much about language.
Remember, the magic word is LINGUISTICS, a very new science, indeed. Morphology is the study of words and their parts: morphemes and lexemes. It examines the grammatical and semantic categories they convey as well as the sounds and sound changes caused by combining them. It also explores the relations of words in sentences.