Table of contents
1.
Introduction to Morphology
2.
Types of Affixes
3.
Classification of Morphemes
4.
Morphological Processes
5.
Morphological Analysis
6.
Applications
7.
FAQs
8.
Key Takeaways
Last Updated: Mar 27, 2024

Computational Morphology

Author Rajkeshav
0 upvote
Career growth poll
Do you think IIT Guwahati certified course can help you in your career?

Introduction to Morphology

Source

In Morphology, we study the internal structure of words and how words are built up from smaller meaningful units called Morphemes. If I take the word 'dogs,' it comprises two morphemes: 'dog' and 's.' 's' is a plural marker of the noun. Similarly, if I take the word 'unladylike,'  it has three morphemes: 'un,' 'lady,' and 'like.' 

Now we will see the various linguistic concepts that we must know before getting started with morphology.  

1. Allomorphs-  Allomorphs are variants of the same morpheme but cannot be replaced by one another, for example, opposite: unhappy,  incomprehensible,  impossible,  irrational. 

2. Bound Morphemes-  Bound morphemes are those morphemes that cannot appear as a word by themselves. For example, we cannot take 'un'  from 'unhappiness' and use it as a single word. This is bound by another word that is applied after it. 

3. Free Morphemes- They are those morphemes that can appear as a word by themselves. For example, 'Happy' is a free morpheme.

4. Stems( roots)- Steam is the core meaning bearing part of the word. For example, in the word 'boys,'  the purpose containing the expression is 'boy.'

5. Affixes-  Affixus are bits and pieces adhering to stem to change their meaning and grammatical functions. For example, we can append 's' after 'boy' to plural.

In general, we can make a correspondence between stems and affixes, bound and free morphemes. We can say that stems are a kind of free morphemes, and affixes are a kind of bound morphemes. 

Types of Affixes

1. Prefix- Prefix is applied before a word. for example, '-un,' '-anti,' '-pra,' etc. (un-happy,  pre-existing).

2. Suffix-  Suffix is applied after a word. For example, '-it,' '-action,' etc. (talk-ing, quick-ly).

3. Infix- Infix is applied between the stem before or after that. For example, in the Philippines,  'b-um-as' is read as 'basa .' Here 'um' is infix. 

4. Circumfixes- Circumfixes are words that precede and follow the stem. For example, in Dutch,  'berg' is 'mountain' and get-berg-te is 'mountains.'

So among these affixes,  the first and second are most common, and the rest are only limited to specific languages. 

Classification of Morphemes

We can classify Morphemes into two categories: Content and Functional morphemes. 

1. Content Morpheme-  Content morphemes are those morphemes that contain some semantic meaning—Example, car, -able, un-, etc.

2.Functional Morphemes- Functional morphemes provide grammatical information about a word.

Example: -s(plural), -s(third singular).

We will generate a new word based on the type of affixes that we applied with the word.  

Case 1- If I have a word walk and apply a suffixing, I will get a new word walking.

Case 2- If I have a word drive with the suffix 'er,' it is converted into the driver.

In the first case, we are not changing the word's grammatical category. 'Walk' is a verb, and it remains a verb in 'Walking.' On the other hand, we started with a verb called drive, which is converted into a noun. So there are two different kinds of morphological processes applied here. The first one is called Inflectional Morphology, and the second one is called Derivational Morphology.

1.Inflectional Morphology- It creates a new form of the same word that is bring => brought,  brings => bringing, etc.

2.Derivational Morphology-  it creates new words by changing part of speech that is logic => logical, illogical => illogical, etc.

Morphological Processes

So far, we have talked about morphemes and different types of exercise that we can apply. But in general, what are the other morphological processes involved to convert one word into another? 

Concatenation- We take various Morphemes and concatenate them to form a single morpheme. For Example,  adding continuous affixes, Hope+less, un+happy, anti+capital+ist+s.

When we combine different morphemes, there may be some changes at the boundary. The changes may happen in the way of the final word pronounced or the final word written. For example, shoe+s => shoes read as shoes.

Suppletion-  In Suppletion, we replace the word entirely with a term that has no connection with the original word. For example, if we convert go into the past tense, it becomes went.

We cannot find any connection between 'go,' and 'went' at the surface level. 

From the NLP  perspective,  how do we process these morphologies? What are the different operations that are done? 

Lemmatization- Given a word, can we identify what is the root word of the Lemma? For example, in the word Saw, the root word is See.

Morphological analysis-  Given a word, can we find the Lemma along with the morphological category of the given word.?

Tagging- Given a word, can we find the actual category of the word? For example, the fundamental type of the word went is 'Past.'

Morphological segmentation- Given a word, we can segment it into different morphemes involved in it. Example: 'de-national-iz-ation.'

Generation- We can take a root word and a particular grammatical category To generate a new buzz.

Morphological Analysis

The goal of the morphological analysis is to take a word as an input and produce the morphological parsed output. For example, if the input word is 'cats,' the morphological parsed result will be 'cat  +N +PL .'The result contains stem and additional information; +N  for noun, +SG  for singular, +PL for plural, +V  for verb, etc.

Applications

1. Text to speech synthesis

2. search and Information retrieval

3. machine translation and  grammar correction

FAQs

1. What is Morphology?

In Morphology, we study the internal structure of words and how words are built up from smaller meaningful units called morphemes.

2. What is a Content morpheme?

Content morphemes contain some semantic meaning—Example, car, -able, un-, etc.

3. What is a Functional morpheme?

Functional morphemes provide grammatical information of a word.Example: -s(plural), -s(third singular).

4. What is Lammenization?

In Lemmatization, we can identify the root word of the Lemma. For example, in the word Saw, the root word is See.

5. What is a Suppletion morphology?

In Suppletion, we replace the word with a word that has no connection with the original word. For example, if we convert go into the past tense, it becomes gone.

Key Takeaways

In this blog, we learned the morphological information that helps us capture the data and different linguistic terms along with the applications. Computational Morphology is an essential part of Natural Language Processing, a domain of Machine Learning. Do Check Restaurant Review Analysis using NLP.

 

Live masterclass