Give an account of Lexical Cohesion in English language.
According to research done at http://www.cs.ucd.ie/staff/jcarthy/home/Lex.html, a text is any passage, whether verbal or written, that forms a unified whole. The word unity in this case refers to cohesion. Cohesion is brought about by back referencing, for example, ‘Mary bought a car. She thought it was elegant.’ ‘She’ in this example refers to Mary, while ‘it’ refers to the car. Hobbs (1982), defines cohesion as relations of meaning that are present in a text. These relations describe the text. Cohesion occurs, therefore when the analysis of some elements in the discourse are reliant on that of another. Examples of cohesion include, reference, substitution, ellipsis, conjunction and lexical cohesion.
In the research found at http://www.cs.ucd.ie/staff/jcarthy/home/Lex.html, lexical cohesion is described as the identification of a lexical item that is in some way related to one that appears before it. Through study of the structure of the vocabulary used, one is able to establish the lexical cohesion that exists in the given text. If lexical cohesion is formed by repeating a lexical item, it is then referred to as reiteration. This can be repetition of the same word or use of a synonym, near-synonym or super ordinate. For example, ‘Tom ate an apple. He loves the fruit.’ In this sentence, the word ‘fruit’ is a super ordinate, which is used to refer back to the subordinate word, ‘apple.’ Another example of lexical cohesion is the ‘part to whole’ relationship e.g. between wheels and bicycles.
When many cohesive elements appear in a variety of sentences, they form a cohesive chain, for example, ‘Mary had cake for dinner. The cake was made up of vanilla. Mary really enjoyed it.’ From the example, the word ‘it’ refers to the cake in the first sentence. It is quite evident too from this example that cohesion is also formed as a result of repetition.
According to Jane Morris of York University and Graeme Hirst of Toronto University, Lexical Cohesion is a consequence of chains of related words that lead to continuity of lexical significance. The chains come up from similarity of various texts, which aid in finding text structure. Ability to establish text structure is very important since it gives one an opportunity to understand the deep meaning of the text itself.
A lexical chain is a series of interrelated words in a passage, found either in short sentences or in the entire passage. Chains do not depend on the grammatical structure of the passage; rather, are words that capture a part of the cohesive structure of the passage. A lexical chain can provide a framework for the ruling of a confusing term and enable clarification of the thought that the term represents. Word Net is one lexical source that may be used in the detection of lexical chains. Therefore, formation of lexical chains aids in identification of key units in a given text or passage, and this is very necessary for document summarization.
The greatest property of cohesion is the ‘quality of unity’. This simply means a text is not just made up of a chain of sentences; rather the sentences have to make sense and basically talk about the same thing. Therefore the sentences function together as a whole hence creating unity of text. The nature in which lexical elements in the chain are distributed throughout the text makes it possible to establish the strength of a lexical chain, which since it encloses a context, conforms to the importance of the textual environment it refers to.
According to Halliday and Hasan (1976), cohesion is a way of getting a set of sentences to ‘hang together as a whole.’ It is therefore, necessary for text unity and eases understanding of the text, without necessarily using dictionaries as well as being able to understand meanings of words without necessarily checking for their meanings. This helps students to learn faster and read more.
Halliday and Hasan (1976) further classified Lexical Cohesion into:-
1. Reiteration; repetition of lexical items, super ordinates, subordinates and synonyms, for example, the word ‘boy,’ could be replaced with ‘lad’, ‘child’ or ‘idiot’, in subsequent sentences. Reiteration is therefore not same as reference, because one does not necessarily repeat the same word in the text. A typical example of this can be cited from Martin Luther King’s speech, ‘I have a dream’, delivered from the steps of the Lincoln Memorial in Washington DC on 28th August 1963. In his pursuit to free the African American, he first refers to the African American as ‘a great American’, and then later in his text he refers to him as a ‘Negro’ and later ‘all men.’
2. Collocation; semantic relationships that occur between similar words, which a community uses to understand texts. It also refers to any occurrence within a text, when lexical items are related to each other in some way, within the language used. They further subdivided these associations into:
Ø Opposites in meaning: e.g. love/hate; man/woman
Ø Pairs of words from the same ordered series: e.g. September/November
Ø Associations between words from unordered lexical sets: Part-whole relations e.g. tyre/car and part-part relations e.g. nose/face and co-hyponymy e.g. blue/green (colour)
Ø Associations based on a history of occurrence e.g. sun, heat, warm etc.
Rafael Salkie on the other hand, is more concerned about the typical relations within a text that form lexical coherence. In response to Halliday and Hasan, other researchers have also discussed lexical cohesion, though it is still widely agreed that, cohesion is the way in which texts are linguistically related. Lexical cohesion cannot exist without sentences. That is, cohesive words should be discussed not only as the defining relations that exist between items, but also expression of those meanings within a text.
Brown and Yule (1983) focus on the relationship between cohesion and the text, and argue that lexical cohesion is not always necessary for texts to create semantic relationships between semantic sentences. For example, ‘The phone is ringing. I am using the bathroom.’ As much as the two sentences form a text, they have no lexical cohesion. This means that a text can exist without lexical cohesion, but lexical cohesion cannot exist without a text. Another example of lack of cohesive ties within a sentence, as argued by Brown and Yule is lack of coherence in a text. A text that has lexical cohesion cannot always create coherence. Cohesion is never enough on its own to create coherence. It is necessary to learn that, ‘cohesion is a manifestation of certain aspects of coherence, and a pointer towards it, rather than its course or necessary results.’ (Cook 1994:34).
The theory of Coherence relations (Hobbs 1978; Hirst 1982; McKeown 1985) is used to describe cohesion. The difference between cohesion and coherent is that cohesion refers to ‘sticking together’ which simply implies that a text hangs together as a whole. The sentences are interdependent on each other for progression of the text. On the other hand coherence refers to ‘making sense.’ This means that there has to be sense in the text.
In a coherent relationship, therefore, there is relationship between clauses and sentences. In contrary, a cohesive relationship exists among elements in a text, hence lexical cohesion. Lexical cohesion is computationally realistic to identify, unlike coherence. For example in Hobbs (1978) example:
1. John can open Bill’s safe.
2. He knows the combination.
From the above example, it’s evident that the identity of coherent relations is interpretative while that of cohesion relations is not. If the relation ‘is about the same thing,’ then coherence exists. In the above example, safe and combination are lexically related, meaning they are about the same thing in some way. This example shows how cohesion can be used to identify sentences or texts that are coherently related. Therefore cohesion and coherent are independent of each other, but help bring unity in a text.
It is important to form lexical chains when interpreting a text. Lexical chains can be formed by first of all determining which words from the text or given passage are to be included in the chain. Halliday and Hasan (1976), the words that occur repetitively such as pronouns, prepositions and verbal auxiliaries should not be considered, as well as high frequency words such as do, good etc.
One must also take into account transitivity when forming lexical chains, hence avoid symmetry. Some lexical chains are stronger than others. The strength of lexical chains is determined by:
1. Reiteration; the more repetition in the text, the stronger then chain
2. Density; a compact chain is stronger than a weak chain
3. Length; a lengthy chain is stronger than a short chain
APPLICATIONS OF LEXICAL COHESION
Cohesive Extractive Summarization: A simple summary of a document can be made through identifying and extracting the most necessary sentences from the text, where the importance of a sentence is measured by the availability of key phrases. This kind of summary often contains outliers, which are sentences that don’t fit with other sentences. Lexical cohesion can therefore, be used to identify and omit these outliers, hence improve the summary quality.
Anaphora Resolution: An ideal text first introduces an entity, such as a corporation, by giving its full name. As the text progresses, the corporation is not mentioned much, but instead, phrases like ‘the company’ or ‘it’ are used to mean the same. Anaphora resolution is then used to identify that these short phrases refer to the same corporation mentioned earlier in the text. One way used in anaphora resolution is creating a chain of lexically cohesive words, hence linking sentences in the given text that talk about the same body.
Improved Speech Recognition: Measuring of lexical cohesion can be used to identify when speech detection software has made errors. The incorrect words usually do not stick together with the rest of the text. This helps minimize errors as well as improve on speech.
Improved Optical Character Recognition: Errors in selective character recognition can also be identified through lack of lexical cohesion.
Improved Machine Translation: Errors in machine translation also require cohesion, even if they may be more cohesive than errors in speech recognition and optical character recognition.
It is therefore clear that lexical cohesion is very important to learners in that, it aids in understanding of the text. Students are able to expose the necessary information that the author hints at, by paying attention to the cohesive ties among words. It also helps one understand the coherence of a text or passage as well as improving one’s vocabulary and writing skills. Ultimately, cohesion is one purpose to form coherence in discourse.
Halliday, M.A.K. and Hasan, R., Cohesion in English, Longman 1976
Peter D. Turney, http://www.apperceptual.com/ml_text_cohesion_apps.html