A Critique of Halliday and Hasan’s Cohesion and Coherence Theory
1 Introduction
The theory of cohesion and coherence is generally considered to be established with the publication of Halliday and Hasan’s Cohesion in English in 1976.  Since then, it has been heatedly discussed by researchers from various perspectives, such as discourse analysis, translation, stylistics etc. Cohesion and coherence, being two basic and key concepts in discourse analysis, have attracted both compliments and critiques. In this essay, the author will first clarify the two concepts by introducing Halliday and Hasan’s cohesion and coherence theory (CCT), and then review other researchers’ study on the two notions, and finally employ evidences to critique the theory.

2 Halliday & Hasan’s CCT
Roman Jacobson, is considered to be the first to study cohesion in text (Crowhurst, 1991: 110), but M. A. K. Halliday is widely believed as the most influential in the field. In 1962, he put forward the concept of cohesion for the first time, turning it into a linguistic terminology from an unprofessional use (Zhu, 1995). As for coherence, it was not a linguistic terminology until the publication of Cohesion in English (Zhang, 2000), which triggers abundant subsequent researches on coherence. Since Halliday and Hasan made a comprehensive and systematic study on the two concepts, the essay will take a close eye on their concepts of cohesion and coherence.

2.1 The concept of cohesion
Halliday and Hasan (1976) state that “the concept of cohesion is a semantic one; it refers to the relationship of meaning that exist within the text, and that defines it as a text” (1976: 4). Additionally, they point out “cohesion occurs where INTERPRETATION of some element in the discourse is dependent on that of another. The one PRESUPPOSES the other, in the sense that it cannot be effectively decoded except by recourse to it. When this happens, a relation of cohesion is set up, and the two elements, the presupposing and the presupposed, are thereby at least potentially integrated into a text” (1976: 4). The central concern of the book is devices of cohesion in English, namely, five types of cohesion devices are identified: reference, substitution, ellipsis, conjunction and lexical cohesion. In Introduction to Functional Grammar, Halliday and Hasan (2004) develop CCT further and make some minor modifications, they regard substitution and ellipsis as “variants of the same type of cohesive relation” (p. 563) and put them into one category. Besides, conjunctions are subdivided in a more abstract level. More than that, Halliday and Hasan continue to study cohesion in another work—Language, Context and Text: Aspects of Language in a Social-semiotic Perspective. In this book, they enlarge the scope of the devices of cohesion, which are classified into two types: structural cohesion and non-structural cohesion. Structural cohesion is composed of parallelism, theme-rheme structure and given-new organization; non-structural cohesion, consists of componential relations and organic relations. Both can be subdivided into grammatical devices and lexical devices. In componential relation, grammatical devices include reference, substitution and ellipsis; lexical devices are general and instantial relations. In organic relation, grammatical devices include conjunctives and adjacency pairs, and lexical devices means continuatives. Halliday and Hasan (1989: 82) summarizes the cohesive devices as follows:







NON-STRUCTURAL COHESION

2.2 The concept of coherence
Apart from taking cohesion as their central concern, Halliday and Hasan investigate coherence, or texture as they prefer to call, in Cohesion in English.  Instead of depicting a clear picture of coherence like cohesion, Halliday and Hasan do not give a clear and direct definition of coherence, but put forward the principles or conditions of coherence when illustrating the relation between cohesion. Their basic concept of coherence can be revealed in following lines.
“The concept of TEXTURE is entirely appropriate to express the property of 'being a text.' A text has texture, and this is what distinguishes it from being something that is not a text. It derives the texture from the fact that it functions as a unity with respect to its environment.” (1976:2)
“The concept of COHESION can therefore be usefully supplemented by that of REGISTER, since the two together effectively define a TEXT. A text is a passage of discourse which is coherent in these two regards: it is coherent with respect to itself, and therefore cohesive. Neither of these two conditions is sufficient without the other, nor does the one by necessity entail the other.” (Halliday & Hasan, 1976: 23)
In addition, Halliday and Hasan (1989) express a similar view: 
“Every text is also a context for itself. A text is characterized by coherence; it hangs together. At any point after the beginning, what has gone before provides the environment for what is coming next. This sets up internal expectations; and these are matched with the expectations referred to earlier, that the listener or reader brings from the external sources, from the context of situation and of culture” (p. 48). 
Based from the above citations, coherence refers to the internal elements of a text, and it is “a combination of semantic configurations of two different kinds: register and cohesion.” (Carrell, 1982: 481). 

2.3 Relationship between cohesion and coherence
As for relationship between cohesion and coherence, Halliday and Hasan (1976: 23) point out: “The concept of COHESION can therefore be usefully supplemented by that of REGISTER, since the two together effectively define a TEXT. A text is a passage of discourse which is coherent in these two regards: it is coherent with respect to itself, and therefore cohesive. Neither of these two conditions is sufficient without the other, nor does the one by necessity entail the other.” By saying this, they claim that cohesion and register are two factors contributing to coherence; a text is not coherent with either of the two.  Although Halliday and Hasan (1976: 229) recognize that there is a distinction between cohesion and coherence, they continue taking cohesive devices as the only factor to construct the coherence of a text.  Additionally, Halliday and Hasan (1989:48) illustrate the relationship as follows:
“An important contribution to coherence comes from COHESION: the set of linguistic resources that every language has (as part of the textual metafunction) for linking one part of a text to another.” Based on the above analysis, a conclusion can be achieved that cohesion is a necessary but not a sufficient condition for coherence, that is, a coherent text should have cohesive devices and a text without cohesive devices is not coherent. Meanwhile, cohesion is only one of the ways to realize coherence, to put it in another way, a text is not necessarily coherent with cohesive devices.

3 Critiques of Halliday and Hasan’s CCT
Halliday and Hasan (1976, 1989), with a central concern of cohesion devices from the perspective of language function, have made great contributions to research on CCT. However, each theory has its room to improve, so does Halliday and Hasan’s theory. With more scholars' further devotion to CCT, whilst taking it as the landmark of study in the field, many of them put forward their criticism, thus seek its development. After the publication of Cohesion in English, researches in this field, mainly in 1980s-2010s, are increasingly done by scholars around the globe. It is surprised to find that Chinese scholars have shown more interest than western scholars in CCT when reviewing the literature. 
Since the publication of Cohesion in English, CCT is developed and enriched by linguists from different perspective: Hoey (1991), Martin (1992), Van Dike (1977), Beaugrade (1981), Widdowson (1978), Fries (1983), Mann and Thompson (1987), Hobbs (1979), Sanford (2005), Brown and Yule (1983), Tannen (1984), Hu (1994), Zhu (2001) and Zhang & Liu (2003) etc..(see Du and Cummins, 2011: 84; Xi, 2010: 140). However, Halliday and Hasan’s CCT is challenged by few scholars. Widdowson (1978), Enkvist (1978), Brown and Yule (1983), Carrell (1982), and Johns (1986) give a direct criticism against the theory (Ghasemi, 2013; Zhu, 1995). Development of and critiques against the theory are sorted out as follows.

3.1 Development of Halliday and Hasan’s CCT
Development of Halliday and Hasan’s CCT is around (1) categorization of cohesive devices and (2) defining coherence from different perspectives. In the west, De Beaugrande and Dressler (1981), and Hoey (1991) dig into devices of cohesion and make their contribution to lexical cohesion. What is worthwhile to mention is in China, a great many researchers showed great enthusiasm to cohesive devices, among them, Hu Zhuanglin, Zhu Yongsheng and Zhang Delu are three representatives. In particular, Hu (1994) and Zhang & Liu (2003) systematically enlarged the scope of cohesive devices and formed their own framework of cohesion, which are considered as the second and third generation of the theory if Halliday and Hasan’s CCT is regarded as the first. 

3.1.1 Development on categorization of cohesive mechanism
De Beaugrande and Dressler (1981) outlined their taxonomy of cohesive devices, including recurrence, partial recurrence, parallelism, paraphrase, pro-forms, ellipsis, tense, aspect and junction. Hoey (1991), proposing in his book Patterns of Lexis in Text: “the study of the greater part of cohesion is the study of lexis, and the study of cohesion in text is, to a considerable degree, the study of patterns of lexis in text” (p. 10), emphasizes lexical cohesion and conducts an intensive study on lexical patterning and ten types of which are introduced.
Based on Halliday and Hasan’s CCT, Hu (1994) amplifies the scope of cohesive devices, transitivity, sound pattern and intonation are for the first time taken as devices of cohesion, what’s more, Hu creatively involves pragmatics in cohesive devices, initiating a new perspective that cohesion should be analyzed at various levels and developed into a multilevel model. His cohesive mechanism is summarized in his book (p. 225) as follows: 

Zhang & Liu (2003) further develop the theory and make breakthrough contributions by proposing cross-type cohesion and assuming interpersonal relations can also create cohesion and coherence, which is different from previous studies taking a central concern on textual and ideational relations.  Their framework of cohesion (p. 29) is presented as follows:

3.1.2 Development on defining coherence from different perspective
As is mentioned in 2.2, coherence is not clearly and directly explicated by Halliday and Hasan. Since 1976, scholars from linguistics, psychology and philosophy are seeking a well-accepted definition of coherence. 
Most scholars strive to define coherence in the framework of linguistics. Van Dijk (1977) studies coherence from the perspective of macrostructure, viewing coherence can be achieved in two levels: local coherence by linear relations between sentences and global coherence by macrostructure. Widdowson (1978) analyses coherence from the perspective of illocutionary act. He defined coherence as “the relationship between the illocutionary acts which propositions, not always overtly linked, are being used to perform” (1978: 28). Danes (1974) and Fries (1983) study coherence in terms of thematic progression. 
With the emergence and development of cognitive linguistics, some scholars begin to examine coherence form the perspective of psychological cognition. Brown and Yule (1983), and Givón (1995) treat coherence as a psychological phenomenon and believe coherence is, rather a semantic concept than a psychological concept in people’s mind. Brown and Yule (1983) assumes coherence “is the reader's (or hearer's) effort to arrive at the writer's (or speaker's) intended meaning in producing a linguistic message’ (224-225), claiming that it is the reader’s (or hearer's) background knowledge that determines coherence of a text. Wang (2005) examines coherence with people’s psychological cognition, and Li (2012) aims to construct a cognitive frame model to explicate textual coherence.
After reviewing studies on coherence from various perspectives, philosophically, Du and Cummins (2011) provides a new angle to explicate coherence. They regard coherence as a normative, subjective and abstract concept which should be analyzed in the domain of philosophy of language thus theorize the concept of coherence, which they believe a solution to solve a chaotic research on coherence.
From the above, it can be concluded that coherence is a complicated concept that consensus so far has not achieved. Furthermore, it involves almost every aspect of communication thus triggers different viewpoints varying from different aspects. Such viewpoints are not mutually exclusive, but complementary to each other.

3.2 Critiques against Halliday and Hasan’s CCT
As is mentioned above, many researches dedicate themselves to extending Halliday and Hasan’s CCT. According to Zhu (1995) and Ghasemi (2013), few (Carrell, 1982; Enkvist, 1978; Brown and Yule, 1983, Johns, 1986 and Widdowson, 1978) challenged the theory and made a direct criticism against the theory. The relationship between cohesion and coherence is the most debated issue in Halliday and Hasan’s CCT thus attracts most severe criticism. Some scholars believe cohesion and coherence are not related while others hold an opposite view.
Wdiddowson (1978), Carrel (1982), Enkivist (1978), Johns (1986) and Brown and Yule (1983) are the supporters who assume cohesion and coherence are not related. Brown and Yule (1983) argue a text can be coherent without cohesive devices because they believe cohesion is a formal concept. They employ a cohesive device “and”, which can convey progressive, adversative or causal relations to present a different view on reference, pointing out that one cohesive word is not corresponding to certain meaning. Widdowson (1978), holding a similar view as Brown and Yule (1983), proves that a text can still be coherent with no cohesion mechanism since he hold cohesion is an illocutionary act and coherence is a perlocutionary act which is determined by its pragmatic function. In the contrast, a text full of devices of cohesion may be odd and incoherent. Thus, cohesion and coherence are irrelevant and in different level. Enkvist (1978) argues that cohesive devices cannot guarantee coherence of a text. He questions the theory by using the following example which is frequently quoted by researchers to criticize CCT: “1 bought a Ford. A car in which President Wilson rode down the Champs Elysées was black. Black English has been widely discussed. The discussions between the presidents ended last week. A week has seven days. Every day I feed my cat. Cats have four legs. The cat is on the mat. Mat has three letters.” The text is full of cohesive devices but it is definitely not coherent. Carrell (1982) and Johns (1986) emphasize reader’s background knowledge when questioning CCT, arguing that there is no significant relation between numbers of cohesive devices and coherence, which is equal to say that the two notions are totally irrelevant.
However, more scholars propose cohesion and coherence are related, holding cohesion is a necessary but not a sufficient condition to coherence. Fries (1983) identifies four elements contributing to coherence: communication purpose, textual meaning, register and cohesion, meaning cohesion is one of the factors to realize coherence. Hu (1994) points out that a cohesive text cannot be defined as a coherent one; nevertheless, a text which appears not to be cohesive tends to be coherent. Zhu (1997) argues that a text with cohesive devices could not guarantee the coherence of a text and cohesion is not the criterion of coherence. However, a coherent text is formally cohesive. 
To sum up, criticism against relationship between cohesion and coherence lies in Halliday and Hasan’s overemphasis on cohesive devices to discourse coherence and their unclear description of coherence. To make clear about the relationship between cohesion and coherence, the core is to define coherence.

4 My Critique of CCT
4.1 Critique of the cohesive mechanism
To critique CCT, it is critical to distinguish cohesion from cohesive devices. In the past decades, researchers have achieved consensus that cohesion is a semantic concept, whereas in terms of cohesive devices there are more divergence than consensus. Working in the framework of Halliday and Hasan (1976, 1989), Hu (1994) and Zhang and Liu (2003) make significant contribution to the theory. Hu (1994) enlarge the scope of cohesive devices to phonological level, which fills in a gap of Halliday and Hasan’s theory and without no doubt a significant supplement to the theory. This is especially the case in poems. Take Alfred Tennyson’s short poem Break, Break, Break for example. 
Break, break, break,
         On thy cold gray stones, O Sea!
And I would that my tongue could utter
         The thoughts that arise in me.

O, well for the fisherman's boy,
         That he shouts with his sister at play!
O, well for the sailor lad,
         That he sings in his boat on the bay!

And the stately ships go on
         To their haven under the hill;
But O for the touch of a vanish'd hand,
         And the sound of a voice that is still!

Break, break, break
         At the foot of thy crags, O Sea!
But the tender grace of a day that is dead
         Will never come back to me.
Cohesive devices in phonological level, according to Hu (1994), incorporate intonation, information structure and sound pattern. Sound pattern includes alliteration, assonance, consonance, rhyme, rhythm and iamb (Hu, 1994, 168). Hu (1994, 168) proposes sound patterns are endowed with cohesive function beyond sentences. Let us examine how sound patterns function as cohesive devices in the poem. Some of the sound patterns are listed in the table below (From my written component of presentation in my literature class HPEN 7003). 



The most obvious one is the consonance in Line 1 and 13, even for readers who don’t know poems can conceive the beginning and the end echo from each other not only in meaning but also in structure. The second is the end rhyme /i:/, /ei/, which display the poet’s intention to make his poem a cohesive one. Assonances as /ei/, /ou/, and /u/ are mentioned at the first stanza, expressing the poet’s mournful emotion to his best friend. The most frequent used sound pattern is alliteration, which runs through the poem from the beginning to the end, playing as cohesive devices thus making the whole poem a coherent one.
Halliday and Hasan (1976, 1989), and Hu (1994) study cohesive mechanism from textual and ideational perspective. Zhang and Liu (2003) extend the scope, proposing interpersonal relations such as mood, modality is contributing to cohesion. Again, let us take Tennyson’s short poem for example. Only one mood adjunct “never” and two modal verbal operators “would”, “could” are used in the poem. “never”, as a negative word, expresses stronger effect than “not”. The Poet’s best friend is dead and this is an objective fact that cannot be changed, so “never” vividly depicts Tennyson’s grief and sadness. The two modal verbal operators “would” and “could”, meaning “willingness” and “probability”, represent Tennyson’s subjective feeling. With the appropriate employment of modality, Tennyson completely conveys his inner attitude and emotions to his friend. By doing so, the poem is cohesive in content, dedicating to express the same theme. After Zhang and Liu (2003), Cheng (2007), based on the Appraisal Theory, constructs an attitude analyzing model and proves that interpersonal meaning has the function of coherence (p. 24). 
To conclude, the paper appreciates Hu (1994)’s contribution to cohesion mechanism studies and advocates taking interpersonal relations into account to extend the scope of studying cohesive mechanism, which is the direction for future studies to further development.

4.2 Critique of the concept of coherence
As is mentioned in 3.1.2, scholars strive to define coherence from different perspectives. Some propose coherence is a semantic concept which exists in the text, others assume coherence is beyond the text (Zhang, 1998). They voice  their views on coherence, but they don’t theorize the concept. The theorization of the concept of coherence, after several decades of study, is high time to accomplish so as to eliminate current chaotic research in coherence. 
Since Halliday and Hasan do not describe a clear image of coherence and previous studies by other scholars are concerned about coherence from different standpoints, which are not exclusive but complementary to each other. Therefore, this paper recommends studying coherence from the perspective of philosophy of language as Du and Cummins (2011) to theorize the concept of coherence. Du and Cummins (2011: 89) argue that cohesion is a scientific concept since it is objective. Cohesion is a concrete concept formed in the objective linguistic world and it is definite; on the contrary, coherence is a philosophical concept and it is a subjective and abstract concept. They also claim that cohesion is given by the writer while coherence is constructed by participants in the discourse, in other words, cohesion is static whereas coherence is dynamic. Therefore, cohesion is monistic rule-governed concept and coherence is dualistic normative concept (Du and Cummins, 2011, 90). In this way, Du and Cummins (2011), different from previous study, taking cohesion and coherence as two sides of coin, regard the two notions in different category thus solve the Enkvist problem. Future studies can be conducted to set up a framework to theorize the concept of coherence.

5 Conclusion
This paper first closely examines Halliday and Hasan’s CCT, and then conducts a comprehensive literature review both in the west and in China, based on that, the paper critiques (1) the notion of cohesion by  employing Tennyson’s short poem to offer some supplements of cohesive mechanism and (2) the concept of coherence by recommending a new perspective to explore what exactly the notion is. In doing so, the paper sorts out researches on the theory, tease out Halliday and Hasan’s contributions to discourse analysis and also shortcomings of their theory, concludes that even after more than four decades of exploration on cohesion and coherence, key issues in the theory are still ambiguous and unresolved, that is, researchers do not have a consensus on what coherence is, which directly leads to a most debated issue: the relationship between cohesion and coherence. To solve the present debate, the key step is to establish universal epistemology on the conception of coherence. In this case, the paper recommends theorizing the concept from the perspective of philosophy of language. Apart from a chaos in coherence, there is room to supplement the framework of cohesive mechanism. Since most previous studies probe cohesive devices from textual and ideational perspective, this paper advocates an interpersonal perspective to improve and perfect current framework.

References
1. Brown, G. & G. Yule. Discourse Analysis [M]. Oxford: Blackwell, 1983.
2. Carrel, L. P. 1982. Cohesion is not coherence [J]. TESOL Quarterly, 16 (4), pp. 479-488.
3. Crowhurst, M. 1991. Interrelationships between reading and writing persuasive discourse [J]. Research in the Teaching of English, 23 (4), pp. 110-122. 
4. Enkvist, N. E. Coherence, pseudo-coherence, and non-coherence [A]. Ostman. J. O. Cohesion and Semantics [C]. Abo, Finland: Abo Akademi Foundation, 1978.
5. Fries, P. On the Status of Theme in English [A]. Petofi, J. & E. Sozer. Micro-and Macro-Connexity of Discourse [C]. Hamburg: Busde, 1983.
6. Ghasemi. M. 2013. An Investigation into the Use of Cohesive Devices in Second Language Writings [J]. Theory and Practice in Language Studies, 3 (9), pp. 1615-1623.
7. Halliday, M. A. K., & Hasan, R. Cohesion in English [M]. London: Longman, 1976.
8. Halliday, M. A. K., & Hasan, R. Language, context, and text: Aspects of language in a social-semiotic perspective. (2nd Ed.). Oxford: Oxford University Press, 1989.
9. Halliday, M. A. K., & C. Matthiessen. An Introduction to Functional Grammar (2nd Ed.). London: Arnold, 2004.
10. Hoey, M. Patterns of lexis in text [M]. Oxford: Oxford University Press, 1991. 
11. Johns, A. M. 1986. Coherence and academic writing: Some definitions and suggestions for teaching [J]. TESOL Quarterly, 20 (2), pp. 247-265. 
12. Van Dijk. T.  Text and Context: Explorations in the Pragmatics of Discourse [M]. London: Longman, 1977.
13. Widdowson. H. Teaching Language as Communication [M]. Oxford: Oxford University Press, 1978.
14. 程微. 2007. 从态度韵看语篇的衔接与连贯[J]. 外语教学 (6): 24-28.
15. 杜世洪, 卡明斯. 连贯是一个语言哲学问题—四十年连贯研究的反思[J]. 外国语 2011(4): 83-92.
16. 李天贤. 2012. 认知框架视角下的语篇连贯研究[D]. 杭州: 浙江大学人文学院.
17. 张德禄, 刘汝山. 语篇连贯与衔接理论的发展及应用[M]. 上海: 上海外语教育出版社, 2003. 
18. 张德禄, 2000, 论语篇连贯[J]. 外语教学与研究 (2): 103-109.
19. 张建理. 1998. 连贯研究概览[J]. 外语教学与研究 (4): 40-46.
20. 朱永生. 1995. 衔接理论的完善与发展[J].外语界 (3): 66-69.
21. 朱永生. 1997. 韩礼德的语篇连贯标准—外界的误解与自身的不足[J]. 外语教学与研究 (2): 20-24.