Some time ago I was asked to assess a new student who was having trouble at school. He misbehaved in all his lessons and was rude to his teachers, meaning that he spent most of his time in detention. Fortunately, staff members wondered if he might have some undiagnosed difficulties that were causing him to act in this way, so he was referred for a Speech and Language Therapy (SALT) assessment with me.
Contrary to the reports I had heard, he was kind and polite one to one, even wishing me a good day at the end of the assessment. But the interview was illuminating. When we talked about how school was going, he told me that he understood “nothing” in his lessons, so there was “no point” in behaving.
I suspected a language disorder, so used the CELF-5 (Clinical Evaluation of Language Fundamentals) standardized assessment to find out more. As expected, he had mild difficulties across a range of areas such as vocabulary and ability to follow instructions. But most concerning was his performance on the Understanding Spoken Paragraphs subtest, when he answered only one question correctly.
The task involves the assessor reading several short paragraphs aloud before asking students questions about them. I like to use it because I think it reflects quite well how much a child may understand in class, but the skills it assesses are quite broad. A low score might indicate difficulties in a number of areas such as with paying attention, auditory memory, vocabulary knowledge, understanding of sentence structures as well as ability to predict and infer.
As I tried to delve deeper, we read the paragraphs again and I asked my student to paraphrase some of the sentences. To my surprise, his sentence level understanding was not bad. Yet he still struggled to say what the paragraph was about, and to answer the text-level questions. What he appeared to be struggling with was understanding how the sentences hung together as a whole, to form a “text”.
As Speech and Language Therapists (SLTs), we are used to supporting our students with some aspects of text level discourse. It is common to work on inference, understanding of different levels of questioning, as well as teaching a range of general reading comprehension strategies such as summarizing, finding the main idea, and perhaps visualizing.
One area that is not so well researched, perhaps even neglected in our field despite being closely linked with comprehension abilities is that of text cohesion. Simply put, this is the “glue” that connects sentences together and distinguishes a unified text from a series of unrelated sentences.
The concept was first introduced and formalized in the seminal work of Halliday and Hasan (1976). According to the authors, for a passage of information to be regarded as a text, certain features must be present. We are talking not just about the organizational macrostructures that characterize different genres (although these do add to the “texture”), but the “cohesive ties” that connect individual sentences to each other grammatically or semantically (Halliday and Hasan, 1976).
Simply put, cohesion occurs when one element of a passage can only be understood in reference to another. In the text, “A man in a blue anorak wandered over Tower Bridge. He stopped to admire the view”, the word “he” can only be understood by referring back to “a man in a blue anorak” in the previous sentence. Cohesion may go in either direction: when you have to refer backwards, as in the previous example, it is called anaphora. This is by far the most common type of cohesive tie. (Halliday and Hasan, 1976).
It is also possible to refer forwards, e.g. “When he got home, John went straight to bed” and “I couldn’t believe she did it: Sarah won the marathon”. In both examples, pronouns are used early on to refer forward to information that comes later. So in the second example, “it” refers forward to “the winning of the marathon”. This is called cataphora. Interestingly, this is the primary function of a colon: as a kind of visual arrow pointing forwards (Halliday and Hasan, 1976).
Halliday and Hasan (1976) classified cohesive devices into 5 categories: reference, substitution, ellipsis, conjunction and lexical cohesion. According to them, the first three are grammatical in nature, whilst the last is semantic, with conjunction being a combination of the two. However, they were quick to note that there is no sharp dividing line between categories and some examples may be on the borderline (Halliday and Hasan, 1976). Other researchers have organized them in different ways.
Reference involves replacing a specific entity in the text with another, as in the man in the blue anorak example. Common reference items include the personal pronouns, “he”, “she”, “they”, “it”, etc, as well as the demonstratives “this” and “that”. The definite article “the” also plays a role, so named because it denotes something definite, or specific.
In the text, “A stray cat appeared on my doorstep this morning. The cat looked hungry, so I gave it some food”, we understand that the cat being mentioned in the second sentence is the same as in the first, since “the” is frequently used to link back to a noun that has already been introduced. Reference mainly involves replacing nouns, though this could be people, animals, places (e.g. there) or objects. In extended reference, whole paragraphs may even be replaced, e.g. “considering all this”.
In substitution, a word or phrase is also replaced with a filler word, but something is added that contrasts with the original idea (Halliday and Hasan, 1976). For example, in the text, “My pen has run out of ink. I need to buy a new one”, “one” substitutes for “pen”. But we are not talking about the exact same pen; we want a new pen. Substitution may be nominal, verbal, or even clausal. Common examples include “one”, “any”, “same”, “do so”, “not”, etc, “Is James coming to the party? I hope so”.
Ellipsis is similar to substitution, but redundant information is removed instead of being replaced. Halliday and Hasan (1976) called it “substitution by zero”. So in the text, “There were ten laptops available in the shop yesterday. Now there are only three”, we can only understand what “three” is referring to by looking back to “laptops” in the previous sentence.
Conjunctions and adverbial conjuncts can also be used as cohesive devices, since they link different propositions together across sentences and make the relationships between them explicit. These relationships may be additive, adversative, causal, temporal, explanatory, or other, e.g. “furthermore”, “despite”, “anyway”, etc.
Finally, sentences may be linked together using vocabulary in lexical cohesion, which can be further subdivided into reiteration and collocation. According to Halliday and Hasan (1976), nominal reiteration is very similar to pronominal reference, grammatically speaking, since the original noun is replaced by something else.
However, instead of a personal pronoun being used, this “something else” may take the form of a synonym, superordinate, subordinate or general noun (or simply the same word repeated). For example, in the text: “The neighbour’s tiny puppy was left outside in the freezing rain all night. The poor thing was shivering uncontrollably when we found it”, “the poor thing” refers back to the puppy.
As well as acting as a cohesive device it allows writers to share more information or a personal feeling about their subject, whilst avoiding repetition. It is a tool relied on heavily by journalists in newspaper articles. This is also one of the examples that falls on the borderline between categories according to Halliday and Hasan (1976), having both a syntactic and semantic element.
With collocation, the linkages between sentences are of a more general nature. Here, we are not focusing on a particular referent. Instead, the presence of related words found throughout the passage tie it together. The more likely the words are to be encountered in the same context, the more cohesive the passage will be. So the words “surgeon”, “craniotomy”, “procedure” and “incision” will give a certain type of “texture”.
Some types of anaphor are easier to understand than others. A simple cohesive tie connects one sentence with the preceding, and is probably the easiest to understand. More remote ties that span several sentences will be more difficult because they tax working memory capacity (Irwin, 1986). Equally, nominal reference is thought to be easier than verbal reference to understand, whilst replacement of whole clauses and paragraphs of information is thought to be the hardest.
In her work on later language development, Marilyn Nippold (1998) observed how understanding and use of a range of cohesive devices continues to develop throughout the school-age and adolescent years. But according to Nelson (1998), this is something that many with Developmental Language Disorder (DLD) struggle with. Indeed, young children with DLD are known for having difficulty with pronominal reference, and inconsistency with this is something I have noticed even at secondary school age.
However, an even greater concern has been some of my students’ difficulties with lexical reiteration. A student with these types of difficulties might easily imagine several different characters in a text where only one is being referred to by different nouns. If a text contains the sentences, “The doctor walked into the emergency room. The surgeon grabbed the scalpel. The specialist began to work”, a struggling student might imagine three different characters. Such a difficulty will clearly have a profound impact on text level comprehension.
To sum up, despite production and comprehension of text cohesion devices being a common area of difficulty for students with DLD, there is currently little information about this in our field compared to others such as education and linguistics. Although the student I mentioned earlier also had broader inferencing difficulties, poor understanding of text cohesion devices, especially lexical reiteration, was a notable problem for him. In future posts I hope to share some practical ways to help students navigate these difficulties.
Notes
Halliday, M. A. K., and Ruqaiya Hasan. Cohesion in English. London: Longman, 1976.
Irwin, Judith Westphal. Understanding and Teaching Cohesion Comprehension. Newark, DE: International Reading Association, 1986
Nelson, N. W. (1998) Childhood Language Disorders in Context: Infancy Through Adolescence. 2nd edn. Boston, MA: Allyn & Bacon.
Nippold, M.A. (1998) Later language development: The school-age and adolescent years. 2nd ed. Austin, TX: Pro-Ed.
Thoughts of a secondary school SLT 