Monthly Archives: January 2017


The longer that I teach academic writing to graduate students, the more time I find myself spending on metadiscourse. Over time, I’ve come to the conclusion that metadiscourse has a bad name—in the sense of a dubious reputation—and an actual bad name. The dubious reputation is presumably connected to both a general suspicion of academic writing and the many instances of laboured prose we have encountered in our careers as academic readers. I’m sure this suspicion is only exacerbated by the fact that the term metadiscourse is a bit of a mouthful. However, this scepticism is deeply unfortunate since thinking about metadiscourse is a natural way to think about our responsibilities as a writer. And, needless to say, thinking more about our writerly responsibilities is crucial for most novice academic writers, making metadiscourse an indispensable topic.

So what is metadiscourse? Simply put, metadiscourse refers to those places in which a writer explicitly acknowledges that they are constructing a text. More specifically, metadiscourse can be defined as “the range of devices writers use to explicitly organize their texts, engage readers, and signal their attitudes to both their material and their audience” (Hyland and Tse, 2004). When we use metadiscourse, we are structuring a three-way relationship between the text, the reader, and the writer. Given our general anxiety about constructing a text that will satisfy the reader, we often neglect our responsibility to be present as the writer of the text. One of my most frequent comments on graduate student writing goes something like this: ‘You are telling me a great deal about your topic but not enough about the text that you are constructing’. This imbalance matters because, as a reader, I need guidance on how to read the text in order to engage fully with the topic.

In my experience, defining metadiscourse is necessary but far from sufficient. That is, a definition of metadiscourse—regardless of whether it is simple or more technical—does little to move graduate students past the sense that metadiscourse is a foreign or artificial textual intervention. To move past this discomfort, I find it helpful to provide a breakdown of different types of metadiscourse and then give examples of each. (For a more detailed version of this breakdown, see Ken Hyland and Polly Tse. “Metadiscourse in Academic Writing: A Reappraisal.” Applied Linguistics 25, no. 2 (June 1, 2004): 156–77.)

In general, we use metadiscourse to signal the following things to our readers:

How our text is organized:

I will start by presenting some of the literature that assesses governmental responses to AIDS in Uganda, Zimbabwe, and South Africa.

Because they share many key concepts, these approaches to the experience of tuberculosis will be organized thematically.

How our ideas relate to one another:

To conclude, the historiography of consumer demand in the eighteenth century has undergone many changes since the inception of consumer studies.

The promotional materials produced by a university often promise that administrators will provide resources to assist students with the transition to university life; as a result, many students arrive on campus with the expectation of support.

How we are using evidence to support what we are saying:

Yet, as the American historian John L. Brook has demonstrated, Habermas’s account of the public sphere seems unable to reconcile the complexities of power.

Swain posits that language learning may occur though the production of language, either spoken or written.

How we are further explaining an idea:

Global norms are norms that are accepted worldwide; for example, it is currently a global norm that all students progress through the degree granting process by completing a series of homework assignments, exams, and research papers.

An assertion of ‘personhood’ expressed as a relation to property is crucial in every self-styled extension of the Enlightenment project. That is, when we equate personhood with property ownership, we implicitly accept a liberal notion of identity.

How much strength we attribute to a particular claim:

Hypothesized reductions in co-rumination during PMT/CBT may also be due, in part, to improvements in the mother’s depressed status.

To my knowledge, this problematic has never been critically examined.

How we feel about a particular aspect of our text: 

This remarkable achievement shows that policy goals are achieved more readily when those policy goals are clearly established.

Understanding the nature of the developments leading up to the unification of Upper and Lower Egypt is particularly problematic because of the lack of literacy in that time period.

How we want readers to orient themselves to a particular aspect of our text:

It is widely recognized that natural resources come to count as such through specific decisions, institutional practices, and socio-political processes.

This claim raises an obvious question about the clinical similarity of patients with aggressive dementia and patients with general stress disorders.

How the text reflects our authorial role:

My use of the term ‘revival’ here stems from an understanding of cultural revitalization as a flexible and organic process, wherein members of a community fuse individual innovation and musical sensibilities with contemporary interpretations of older cultural practices.

Thus, I will try to link insights from theoretical understandings of science and technology studies with resources geography, which may potentially advance both these literatures.

Examples are taken from student writing and used with permission.

What we see in these examples is how naturally most of them read. Rather than seeming stilted or artificial, these sentences appear to be doing important work. In fact, if we were to return to the broader passages from which I extracted these sentences, we could see that those passages work better with these sentences than they would without them.

The reason that I think it is helpful to consider this breakdown is that a typology allows us to see that we may have very different patterns of use for different types of metadiscourse. To deepen our ability to use metadiscourse well, it is essential to understand these patterns. One of the biggest obstacles to using metadiscourse effectively is a tendency to see it as synonymous with signposting and to imagine that all signposting has to be clunky and awkward. To help students see how they might be able to use metadiscourse better, I like to divide usage patterns into four basic types:

  1. We may use some kinds of metadiscourse pretty routinely; for example, most academic writers use evidence frequently, if not always effectively.
  2. We may avoid some kinds of metadiscourse instinctively because we believe that they may violate the norms of academic writing; for example, some academic writers avoid the first person or affective language that could signal their attitudes or how they wish readers to see the text.
  3. We may use some kinds of metadiscourse hesitantly or inexpertly due to inexperience with academic writing; for example, some academic writers may struggle to provide clear transitions or explanations and may have difficulty identifying the appropriate strength for their claims.
  4. Finally, we may under-use some kinds of metadiscourse because they require an understanding of our own texts that we lack; for example, many academic writers fail to explain the structure of their own text adequately because they don’t yet understand its internal dynamics properly.

These different orientations show us the fundamental inaccuracy of any attempt to see metadiscourse as good or bad. As writers, we can use these four categories to develop questions that will challenge our own writing practice:

  • If we are using some sorts of metadiscourse routinely, are we doing it well?
  • If we are avoiding some sorts of metadiscourse, can we deepen our understanding of the norms of academic writing to be sure this judgement is based on a sound understanding of disciplinary practice?
  • If we are using some sorts of metadiscourse hesitantly or inexpertly, could we improve our understanding of the value of such devices for the reader and thus overcome our reservations?
  • Lastly, if we are under-using some sorts of metadiscourse because we lack a sufficiently deep understanding of our own text, can we learn how to develop that understanding in order to meet the reader’s need for guidance through our text?

Taken together, these questions can help us to see how we might adjust and thus improve our use of these many varieties of metadiscourse. I recommend that any academic writer devote some revision time to the identification of the metadiscourse employed in their own texts. My strategy for doing this would be to keep in mind the breadth of work that metadiscourse accomplishes without focusing too much on the sort of classification found in this post. For most writers, it is sufficient to think about all the things we do to guide and engage our readers and look for those. Highlighting those places where we are present in our texts can be hugely instructive for all writers. In particular, if a supervisor is asking about voice or questioning overall coherence, I would use this highlighting strategy to see where you may still be absent in your own text. Even if you are more comfortable with the use of metadiscourse, I would still suggest this highlighting strategy as part of late-stage revision. It is only by coming to an understanding of our role in our own text that we can ensure that our readers will have the guidance that they need to get the most out of our writing.