Although linguists like to describe our knowledge of words as a mental lexicon, we know that the mental lexicon is not organized alphabetically like a dictionary. Rather, it is a complex web composed of stored items (morphemes, words, idiomatic phrases) that may be related to each other by the sounds that form them and by their meanings. Along with these stored items we also have rules that allow us to combine morphemes in different ways. Our evidence for this organization comes from experiments using both normal subjects and subjects with some sort of genetic disorder or trauma to the brain.

There is a great deal of evidence to support the idea that speakers do not merely learn and store complex words (although they may store some complex words which are used frequently), but rather construct complex words using rules of word formation . We will go into great detail to come on exactly what these rules of word formation look like, but let us start with a simple example, and use that example to explore what linguist Steven Pinker calls the “words and rules” theory of the mental lexicon (Pinker 1999).

We will take as our example the rule for forming past tenses of verbs in English. At this point, if I asked you how to form the past tense of a verb in English, you would probably say that you usually add an -ed. And then you might point out that there are a number of verbs that have irregular past tenses like sing~sang, tell~told, win~won, fly~flew, and the like. We will look first at the regular past tense rule.

While it is true that in writing we add an -ed to form the past tense of a verb, in terms of spoken speech, the situation is a bit more complicated.

You pronounce the past tenses of the first set of words in the Challenge box with a [t] sound, in the second with a sound like [əd], and the third with a [d] sound.

We do not choose the pronunciation of the past tense at random. Rather, the choice of which of the three endings to use depends on the final sound of the verb. Those words that are pronounced with final [t] or [d] sounds – those in the second list – get the [əd] pronunciation. The words that end in voiceless (with the exception of [t]) sounds get the [t] pronunciation. And all the rest get the [d] pronunciation. As for irregular forms like sang and flew, we must assume that English speakers simply learn them as exceptions.

We know that speakers of English have an unconscious knowledge of the past tense rule because we can automatically create the past tense of novel verbs. For example, if I coin a verb blick, you know that the past tense morpheme is pronounced [t]. Similarly, the novel verb flurd will have the past tense [əd], and the verb zove will be made past tense with [d]. We can even form the past tense of verbs that contain final sounds that do not occur at all in English, and when we do, we still follow the rule. For example, if we imagine that there are many composers imitating the style of Johann Sebastian Bach, and we coin the verb to bach to denote the action of imitating Bach, we will automatically form the past tense with the past tense variant pronounced [t], because the final sound of Bach is [x], a voiceless velar fricative. The important point here is that when we hear this sound at the end of a verb we know (unconsciously) that it’s voiceless, and apply the past tense rule to it in the usual way

Now that we know something about the English past tense rule, we can return to the question of how the mental lexicon is organized. It might be plausible to assume that speakers of English use the past tense rule when they are creating the past tenses of novel verbs, but simply store the past tense forms of words they have already heard. In other words, we might assume that once a past tense has been formed, it is entered whole in our mental lexicon, and we retrieve it whole just as we would the present tense form. This hypothesis, however, may not be correct.