Consider the data in (17):

Generally, as the examples in (17a) suggest, it seems possible in English to derive an -ity noun from an adjective that ends in the suffix -ous. But in some cases – for example, those in (17b) – the -ity form just does not seem possible. Why should this be? The examples in (17) illustrate a phenomenon called blocking. Blocking occurs when there is another word that bears the same meaning or fulfills the same function as the non-existent word. In the case of the examples in (17b), but not (17a), there are simple, underived nouns from which the -ous adjectives are formed: glory, fury, and grace. Since those words exist in English, they block or preclude the formation of the -ity nouns. The -ous adjectives in (17a) do not derive from free morphemes (they are built on bound bases), and therefore can form nouns by affixation of -ity.

Blocking can also be seen in nominalized verbs in English:

As the examples in (18) show there are a number of different suffixes that form nouns from verbs, among them -tion, -ance, and -ment. Each verb seems to choose one affix with which its nominalization is derived. Other nominalizing affixes cannot attach. It appears that the existence of one nominalization blocks the existence of others.

Blocking can be seen in inflectional paradigms as well as in cases of derivation. For example, in English the suppletive past tense form went blocks the formation of a regular past tense form *goed, and the existence of the irregular plural form children blocks the formation of the irregular plural *childs.

The obvious question that a theorist might ask is why blocking occurs. One possible answer is that languages tend to avoid synonymy. Once we have a word that means ‘more than one child’ or ‘the process or result of reserving’, why would we need another? Evidence that supports this hypothesis in fact comes from the relatively rare examples where we do find ‘doublets’ – bases that do take more than one plural or nominalizing affix. Consider the examples in (19):

Superficially, these examples seem to provide evidence against blocking. In (19a) we have two plurals of the word brother, one the regular plural brothers and the other an archaic plural brethren. In (19b) we can see that the verb commit has three different nominalizations, not just one. But these examples do not really argue against blocking, because we don’t in fact have cases of synonymous forms. Although brethren was at one time just a plural of brother, it has specialized in meaning and is used in religious contexts to refer to members of the same church. In the case of the different nominalizations of the verb commit, each one has a specific lexicalized meaning. Commission, for example, is the act of committing, and a commission an order to create a piece of art. A commital is an order to send someone to prison or to the hospital. And a commitment is a pledge of certain sorts. Blocking doesn’t occur in these cases because we do not have words that are synonymous.

Our hypothesis that blocking is the result of the tendency to avoid synonymy would therefore seem to be a strong one. Nevertheless, there are other data that call it into question. Consider the examples in (20):

As the examples in (20) show, alongside a noun formed with -ity it is always possible to form a noun with -ness. Although occasionally the two words have distinct meanings (for example monstrosity and monstrousness mean different things), much of the time the -ity and -ness words do appear to be synonymous, contrary to our hypothesis. It is therefore not possible to say that blocking always occurs to avoid synonymy, although we can say that this is a clear tendency in languages. One reason for the ability of -ness to form nouns alongside -ity is that -ness is so very productive in English. We might therefore say that blocking can only be overridden by the most productive of affixes.