The Problem

So far we've only talked about morphological structures in a static sense. We say words 'have' affixes and are 'consist of' morphemes. However, it is often more convenient to describe a complex word as an outcome of a process. For example, we can say that affixes are attached to the base or complex words can be derived¹ from simpler ones. Thus, we need an approach to formally describe these processes.

In this article, we will first explore various morphological patterns from different languages before introducing two approaches for describing morphological rules.

Morphological Patterns

We have seen examples in English like $\small cat-cats,\ book-books$. In English, one way to form the plural of a noun is by adding a $\small \textit{-s}$ suffix. So it is clear that $\small \textit{cat-s}$ consists of two morphemes.

However, consider the following examples in German:

One way to form plurals in German is by changing the vowel in the stem, where the plural forms cannot be segmented into two morphemes. In cases like this, it is more intuitive to think of something has changed rather than something has been added.

Then, a morpheme is just a special subtype of morphological patterns.

Linguists often distinguish two basic types of morphological patterns:

• concatenative: when two morphemes are ordered one after another
• non-concatenative: everything else except concatenative ones

Affixation and Compounding

Affixation is the most common concatenative morphological pattern we have seen so far. We use affixation (subtypes including suffixation, prefixation, etc) and compounding.

Combinatory Potential

It is not the case that any affix and any base can combine. For example, we can combine $\small un-$ and $intelligent$ via affixation to form $unintelligent$, but $intelligent$ and the suffix $-able$ cannot be combined to create a new word.

A rule of affixation is also a statement about which types of morphemes may combine, which is the combinatory potential. Terms like subcategorization frame and selectional restriction are widely used as well.

The combinatory potential of an affix cannot be entirely predicted from its meaning. The word-class of the base is an important factor for combinatory potential. For example, The prefix $\small non-$ and $\small un-$ are similar in meaning, but $\small non-$ is commonly attached to nouns but less readily to adjectives.

The combinatory potential of the prefix $un-$ can be expressed with the notation $\small [-\text{A}]$, where $\small '-'$ stands for the affix and $\small '\text{A}'$ indicates both the word-class of the base and the position of the base relative to the affix.

Base Modification

We have previously defined 'base' as the part of a word that suffixes attached to, but this definition does not apply to patterns such as the formation of German plural nouns, thus forcing us to revise out original definition of base.

The base of a morphologically complex word is the element to which a morphological operation applies.

Base modification is one important class of non-concatenative morphological patterns. It is a collective term for morphological patterns in which the shape of the base is changed without adding segmentable material.

Common types of base modification pattern include:

• palatalization
• fronting
• weakening
• gemination (consonant lengthening)
• lengthening
• shortening
• tonal change
• stress shift
• voicing
• substraction (delete one or more segments from the base)
• metathesis (switching of two or more segments within the base)

Non-concatenative morphological processes are similar to concatenative processes in having restrictions that are equivalent to combinatory potential.

Reduplication

Reduplication is a common morphological operation, whereby part of the base or the complete base is copied and attached to the base.

The element attached to the base often consists of both copied segments and fixed segments, so this kind of mixture between affix and reduplicant result is called a duplifix.

Examples to be Added

Conversion

Conversion is a limiting case of morphological patterns, where the form of the base remains unaltered.

Conversion is generally invoked only for derivational morphology, and primarily for relating two lexemes that differ only in lexical class.

Outside the Realm of Morphology

There are still many other ways to create new words:

• acronyms: $\small \textit{NASA},\ \textit{radar}$
• alphabetisms: $\small \textit{CD},\ \textit{Ph.D.}$
• clippings: $\small \textit{fridge}$ from $\small \textit{refrigerator}$
• blends: $\small \textit{smog}$ from $\small \textit{smoke}$ and $\small \textit{fog}$

They do not fall under morphology, because the resulting new words do not have different meanings to the longer words from which they are formed.

Two Approaches to Morphological Rules

Morpheme-Based Model

Word-Based Model

TBD