Students understanding the complexity of language

Chapter 12.3: Word Formation by Derivation


chapter 12.3: word formation by derivation

The joining of morphemes together into “words” follows certain rules and patterns (i.e., it is another demonstration of how English works as a system). As native English speakers we all have an intrinsic understanding of these rules as part of our linguistic competence. On this page we will discuss how words can be formed by the adding of derivational morphemes, that is, prefixes and suffixes.

The Hierarchical Structure of Morphemes

The derivational morphemes discussed in 5.1 have specific rules governing how they may be added to words. Consider again the example unlucky. The root morpheme, or base morpheme, in this word is luck, a noun. In other words, it is the root or base to which affixes may be added. If we try to add the prefix un– to it we will realize that it does not “sound right” (using our notions of “acceptability”). Thus, *unluck does not make an acceptable word. The reason for this is that the prefix un– is usually only added to adjectives as a way to negate their meaning. Consider the following examples: unhappy, unable, unused, uncouth, untold. Compare this with the results of adding un– to nouns: *unsofa, *uncar, *uncomputer. None of these form acceptable English words.

(Note: there is a second un– prefix that has a separate origin, meaning “to reverse an action,” that is found before verbs, (unbind, unlock, uncover, undo); this prefix is often confused with the adjectival one and has been so for centuries).

To return to unlucky, we see that in the order of progression for adding bound morphemes, we start with the base morpheme luck, and we first must make this into an adjective, by adding –y. This is a very common “adjectival suffix,” that is, it creates adjectives when it is added to a word (usually nouns, but sometimes also verbs or other adjectives). Lucky therefore means “characterized by luck” or “full of luck.” Once we have the adjective lucky, we can negate it by adding the prefix un-. This process shows that there is a hierarchy or order in which bound morphemes may be added:

Hierarchy of Morphemes in Unlucky

Productive and Unproductive Affixes

The derivational morphemes like un- and -y are Germanic in origin, and so have been part of English since the English was first spoken. We can still use them today to create new words that have not been spoken before. Thus, they are called productive affixes. This means that they are still part of the lexicon that we understand and we consider the use of new words formed from them as acceptable (providing they follow the proper hierarchy). There are a large number of productive affixes: –ness, -ation, -ify, -ful, -age, -ish, and so on.

We have unproductive affixes as well. These are affixes that were once used in English but which we no longer use to form new words, simply because we no longer recognize them as having a semantic meaning. Consider the words strength, health, width, wealth, stealth, length, and breadth. They all end in the suffix –th. This is a suffix that was once used to make nouns out of adjectives (that is, it was a nominal suffix). Compare those nouns with the following adjectives (and one verb) from which they were formed: strong, hale, wide, weal, steal, long, broad. In spite of the existence of these nouns ending in –th, we cannot create new nouns from adjectives by adding this suffix. For example, we cannot make a new noun *wrongth (or more properly, *wrength) to mean “the state of being wrong” in the way that strength means “the state of being strong.”

Here is a list of English unproductive suffixes. You might want to look over this list so that you can see what the rules of adding these suffixes were, and try
to create new words with them. Maybe you can make them productive again.

A side note on the notion of “rules” and of “unproductive” as I am using these terms. When I say that un– can only be added to adjectives, I do not mean this as a prescriptive rule but as a descriptive rule. The only reason that we cannot add un– to nouns is that it does not sound “acceptable” to speakers of English. In fact, this unacceptability can even be exploited by companies looking to call attention to their products. Pepsi tried to distinguish 7-Up from other types of cola by calling it the Uncola, a word that, by not sounding quite right, calls attention to itself and thus the uniqueness of the product. Mercedes Benz tried to do the same for the Smart car, calling it an uncar. Interestingly, they also use the phrase unbig, another word which does not sound acceptable, which shows that there are other rules to the use of un– than simply adding it to adjectives. For example, it never occurs with adjectives describing
colors or sizes (*unblue, *unsmall) or with adjectives that express approval or disapproval (ungood, *unbad). Note George Orwell’s jarring use of ungood to mean “bad” and unlight to mean “dark” in 1984, which imagines a new form of language called Newspeak in which other such oddly formed words are used.

In spite of the usage of such terms in ads or literature, they still are not “acceptable” as part of everyday English. If one of them were to catch on, however, then the rules of acceptability would change.

The same principle of acceptability governs the notion of which affixes are productive and which aren’t. Affixes are unproductive simply because they are no longer used. If we started using them again, they would be considered productive. There is nothing inherent in the affix that makes it unsuitable for productivity.


Recently created affixes

Not all affixes are old. There are a number of fairly new affixes that are in quite common use. Perhaps the most common now is e– (for “electronic”), found on e-mail, e-trade, e-commerce, e-book, and so on. Other new ones are i-, –thon, –gate, –lympics, –mageddon. (On these last four, see under portmanteau).


We have talked about prefixes and suffixes but have said nothing about infixes, which are rare in English. There are several unproductive infixes that were used in Proto-Indo-European or Proto-Germanic that still survive in English but are completely unrecognized as such now. For example, the n in the word stand is an infix, as you can tell by its absence in the past tense form stood. Its meaning apparently was as part of a present-tense marker (if you know Latin, you can find other examples of this same infix, such as present tense fundo “I pour” and past tense fudi, “I poured,” without the -n-).

There are only a few productive infixes in Present-Day English. The most common one uses the word “fucking” as the infix, which you may prefer not to use or to substitute with some phonetically similar combination such as “fricking.” Consider the words “abso-fucking-lutely,” “un-fucking-believable” and “fan-fucking-tastic.” This infix is an intensifying morpheme, and is often called “expletive infixation.” There are tamer versions of these such as “absoposilutely” with -posi- (from positive) substituted, and “fan-damn-tastic” but these are much less common.

Another infix comes in the form of –iz(n)-, originally in hip-hop and rap lyrics: h-iz-ouse, sh-izn-it, etc. These infixes might be considered prosodic morphemes since they change the rhythm of the words (useful in song lyrics), but they also add specific cultural coloring to the words. The use of this infix was popular for about a decade, but seems to have died out recently.

Finally, there is the so-called “Homeric infixation,” not named after the great Greek poet Homer, but after Homer Simpson. The Homeric infix is the -ma- in words like “edu-ma-cated” and “sophisti-ma-cation.” This infix gives words a pseudo-intellectual coloring.

(Note: Some linguists would say that none of the above are true infixes. The expletive infixes fail the test since they are free morphemes, not bound, derivational morphemes like the other affixes discussed here. The hip-hop and Homeric infixes fail since they do not change the meaning of the word in the way that derivational morphemes should. While these are fair objections, I do not feel like they are enough to disqualify them from being considered at least a type of infix).