Mythbuster #2: There are a very small number of categories of words.

We learned in the previous section that words belong to categories on each different dimension. How many categories are there? Are there the same number of categories on each dimension? The answer to the first question is not yet established, but it is clear that the number of categories differs on each dimension. Let’s look at morphology. Here a word can either be a complete word, unalterable in form, like it. We can say it is inert. Then there are roots, which are the basic forms that we can add other things to. For example, we can take the word appeal and add a suffix -ing to get appealing. We can use appealing as a full word, but it is also a stem because we can add another suffix -ly to get appealingly or a prefix -un to create unappealing or even put it all together unappealingly. An affix is something we add on to a word. English has three types of affixes: prefixes, which go in front of words, suffixes, which go at the end, and infixes, which go in the middle of words. The infixes in English are mostly taboo words, and we’ll discuss them in detail in another mythbuster, but if you can’t think of anything, you may find the whole topic fan-*******-tastic. So morphology has the categories: inert, root, stem, and three types of affix. There are also different types of roots. The parts of speech crowd confuses us by referring to them as noun-stem, verb-stems etc. It is true that there is a strong correlation between categories of morphology and those of syntax, but there are also plenty of exceptions. The links are sometimes not with syntax, but with logical structure. Consider the suffix -ish as in reddish, tallish, roundish. It would seem that the suffix attaches to adjectives. We can add it to nouns, too: clownish, foolish. Even proper names and abbreviations can be suffixed in this way: A Reaganish tax cult, An MTV-ish music video, A Jordanish jump shot. It seems that we can suffix -ish to anything which is a logical property or something. This can include a wide variety of grammatical forms, including nouns, adjectives and noun phrases. Indeed, not all nouns can take -ish, for example chili-ish, sandish, and atomish sound decidedly odd to me. I wouldn’t say such forms are impossible, but they are by no means normal. In the mythbuster on grammatical judgments we’ll get into detail about the importance, or non-importance, of odd-sounding forms.

There are fewer categories in logical structure. There are three types of categories. A property is something which can be used to assign certain characteristics to something. If we talk about an electric guitar we are talking about something which has the property of being a certain type of stringed musical instrument which also has the property of being powered by electricity. Note that electric is an adjective and guitar is a noun. Don’t think that most property words are adjectives. There are just as many nouns which are properties, including most common nouns. An object is a usually a combination of a property with a word known as a quantifier, such as a, two, this, those, all, some, none. So an electric guitar is an object. Things commonly called pronouns (another myth we’ll examine later) are objects all by themselves, and so are proper names. A predicate combines with objects. The word play is a two-place, or transitive predicate, which requires two objects to make a coherent phrase. It combines with an electric guitar to the one-place predicate we get play an electric guitar which needs one more object to make sense. If we add another object, Jerry we have Jerry play an electric guitar, which is comprehensible, but not a sentence of English, since we have some rules to follow, such as every sentence in English must have a sense of time encoded. We can indicate past tense by adding a morphological suffix -ed to the verb (though this does not affect the syntax). We now have Jerry played an electric guitar. The past tense is an example of what we call an operator. Other operators include negation (not) and most adverbs. When we have all the necessary elements, including an operator, a predicate and as many objects as are required, we have a complete formula. So logical structure is not especially complicated.

Turning to syntax, we find a much more complex situation. We need to account for both individual words and phrases. It is relatively uncontroversial that there are four major types of syntactic categories: noun, verb, adjective, preposition and then there are a lot of minor categories including adverbs and particles. Each of these can be expanded to make a phrase: Consider the sentence: Jerry played an electric guitar in a band which was fond of jamming.

In addition, there are other levels. One lies between noun and noun phrase, consisting of a noun and adjective but no determiner: electric guitar. Although this level is neglected in most courses on English, it is very important. We can substitute a special pronoun one for just this category: If you think that the show was good, you should have seen the ones they played twenty years ago!

We can simplify the discussion of syntactic categories considerably if we start with a few simple rules:

1. There are three levels of phrases. 0 = single word, 1= intermediate category, 2=complete phrase
1. There are four categories that can act as the foundation (head) of a category
2. Categories without foundations (heads) must combine with other categories.

That gives us at least 12 defined categories (4 heads times three levels) plus the headless (adverb and particle) ones.

These syntactic categories are quite different from those we found in other dimensions. Though we will go into more detail later, this should be enough to illustrate our point that each dimension is different, not only in the areas of grammar they represent, but also in the number and nature of the categories that can be found there. Therefore we cannot talk about parts of speech as it there was real substance behind the notion. Each word must be specified on each dimension, and only then can we try to draw general conclusions about the way things work.