Setting the Story Straight Part 2: Put the history horse before the genre cart

In my previous post, I argued that an exam board recommending one explanatory mode for all types of causation questions is problematic if one accepts that academic history is the ‘benchmark’ for ascertaining what ‘good’ historical explanation is.

My intention in that post was not to argue that the narrative mode is necessarily better. In fact, among theorists of history the debate regarding the inherent superiority of the ‘narrative’ versus ‘analytic’ mode or vice versa has largely fizzled out due to a recognition that any one explanatory model cannot cater for all possible historical causal questions. Instead, the debate has been replaced with a kind of pragmatic ecumenicism. For example, Stone argued that;

History has always had many mansions, and must continue to do so if it is to flourish in the future. The triumph of any one genre or school eventually always leads to narrow sectarianism, narcissism, and self-adulation, contempt or tyranny towards outsiders, and other disagreeable and self-defeating characteristics. [i]

Accordingly, some theorists of history recommend adapting one’s choice of mode depending on the question posed. Atkinson, for example, argued that ‘the truth surely is that some questions require narrative, others analytic answers’.[ii] The question discussed in my previous post, for instance, invites the student to explain an individual’s action across a relatively short time-span;

How far does the role of Governor Phips explain the end of the Salem witch hunt (1692–93)?[iii]

Such a question might be more satisfactorily explained in the narrative mode organised chronologically. Other questions from the exam board, however, require the student to focus explicitly on preconditions. For example;

‘It was the unusual political conditions operating in Massachusetts in 1692 that explain the extraordinary events in Salem.’

How far do you agree with this explanation of the Salem witch-hunt of 1692?[iv]

This type of question, by contrast, might be better explained with an analytic response. The student could organise their paragraphs in the form of largely slower-moving, abstracted ‘factors’ structured rhetorically in terms of relative importance.

Possibly more importantly, when faced with more open-ended causal questions there needs to be, in my view, some recognition that one’s choice of mode is itself argumentative. For example, in their guidance for undergraduate history students the historians Marius and Page suggested one should;

keep in mind that argument, in the sense of developing a thesis, is fundamental to all modes used in writing history essays. The modes overlap, and you may use all of them in a single essay; certainly we have in our own writing…when you write an essay, try to determine which modes will best advance your argument.[v]

For example, let’s imagine a student was presented with the question;

            Why was there a witch craze in Salem in 1692?

The student wishing to emphasise the causal importance of individual action over a short time period could adopt the narrative mode. By contrast, another student seeking to argue that structural preconditions were more important might opt for the analytic mode. Alternatively, the student looking to provide an explanation incorporating both might shuttle between the two modes. In fact, most historians now tend to recognise the importance of fairly representing structural conditions, individual actions, and their interplay. Tosh, for instance, noted the importance of such argumentative flexibility;

The truth is that historians need to write in ways that do justice to both the manifest and the latent, both profound forces and surface events. And in practice this requires a flexible use of both analytical and narrative modes: sometimes in alternating sections, sometimes more completely fused throughout the text. This, in fact, is the way in which most academic historical writing is carried out today.[vi]

There has been an understandable delay as trends in academic history have been ‘recontextualised’ into school history. History teachers have built on seminal works such as Byrom’s emphasis in 1998 of the huge intellectual challenge of constructing a historical narrative from an incomplete and contradictory evidence base and Lang’s defence of narrative in history curriculum design in 2003.[vii] The cumulative power of the history-teacher curricular theorisation that has followed in their wake has been evidenced in the National Curriculum Key Stage 3 reforms which speak of the importance of the narrative mode and the new GCSEs in which narrative questions are now set.[viii]

I would argue, however, that for at least one exam board certain generic genre-based assumptions still exist that actively discourage students from writing historically. For example;

  • At GCSE, there are commands such as ‘Describe…’, ‘Write a narrative account…’ or ‘Explain…’ which put the genre cart before the historical horse.[ix]


  • The command ‘Explain…’ is applied in its generic and everyday sense where it can have two meanings: ‘Explicate’/’Explain what’ or ‘Explain why’. This means the command is indiscriminately given irrespective of the second-order conceptual focus of the question. [x]


  • At GCSE, such commands artificially silo students’ writing into discrete genres such as ‘Description’, ‘Narrative, and ‘Explanation’.[xi]


  • At GCSE, ‘Description’ and ‘Narrative’ genres are still caricatured as ‘lower-order’ in the sense that these question-types are typically assigned fewer marks. [xii]


  • At A-Level, as I discussed in my previous post, some examiners still actively penalise narrative irrespective of the type of question asked.


If we take academic history as our benchmark, however, we see a rather different picture.

  • As I discussed above, theorists of history have increasingly recognised that the history drives the genre. The genre shouldn’t drive the history. For example, Goldstein’s frustration regarding the strict prescriptions of modes over forty years ago is perhaps as relevant to discussions in history curriculum design today;

It cannot be reasonable that the essential nature of the discipline is defined by the literary form in which its results are conveyed rather than by the kind of inquiry it is. That so many serious writers think that that is exactly how history is to be characterised is precisely the mark of how far conceptual analysis and the elucidation of modes of discourse have managed to submerge real epistemology.[xiii]

  •  As I discussed in a previous postin history ‘explain’ usually means ‘explain why’ not ‘explain what’/’explicate’.


  • As I discussed in my previous post where I explored how in certain circumstances narrative/description can be highly explanatory, in history genres are conflated. As Megill suggested, for example, ‘narrative blends description and explanation’. [xiv]


  • Historical writing, irrespective of mode, is argumentative. Or, to put it another way, if the mode isn’t argumentative, it’s not history.[xv] For example, Fischer argued that regardless of the mode a historian chooses to write in history is ‘always’ ‘articulated in the form of a reasoned argument’.[xvi]

So, if a question asks students to ‘describe two features of x’ (2 marks), this is ‘lower-order’ only in the sense that a) students are misleadingly expected to treat the object of description ‘x’ as unequivocal and b) there is no overarching historical question in which the student is supposed to situate their description. It’s easier precisely because it’s a non-argumentative, non-historical question. This type of question hardly reflects the argumentative properties of historical description per se.[xvii]

  • Because all modes of history are argumentative in the sense that claims have to be justified through evidence and contribute to the historian’s overall thesis, it follows that historical description/narrative is not easy. For example, one of the mode’s chief defenders Elton argued narrative;

is not, of course, easy to write. It challenges the constructive skills of the writer as well as his ability to hold together a great number of traces that keep pulling apart. It is much harder to do than analytical history.[xviii]

Hewtison, building on Walsh, suggested that one common misconception that underpins the denigration of narrative is the failure to distinguish between its  ‘plain’ (non-historical) and ‘significant’ (historical) guises.[xix] A ‘plain’ narrative only aims to chronicle what happened and is not intended to argue in response to an overarching question. ‘Significant’ narratives, by contrast, are harder precisely because they are problem-oriented.[xx]

Similarly, Megill suggested that historical description also presents ‘an infinite number of difficulties and possibilities’.[xxi] Even the most seemingly innocuous descriptive claim needs to withstand critique from the probative community of historical scholars. Take, for example, a statement such as;

‘The Salem witch crisis took place in 1692’.

Straightforward description right? Well, not quite. In history, where one has to argue using fragmentary and discrepant evidence when describing extremely complex social phenomena in the past, seemingly inarguable statements such as this can themselves be the source of historical controversies.

For example, both Norton and Baker have argued that focusing on ‘Salem’ fails to do justice to the known evidence in the sense that it downplays the geographical reach of accusations which spread to twenty-two separate towns or fails to recognise that more accusations and confessions took place in neighbouring Andover. Accordingly, they argue that a Salem-centric description is misleading and that ‘Essex County witchcraft crisis’ or ‘New England witchcraft crisis’ would be more appropriate.[xxii]

And what about ‘crisis’? In his recent historiographical overview of the Salem witch trials Fels criticised historians who chose ‘crisis’ rather than ‘hunt’ by arguing that the former privileges the role of impersonal forces and downplays the agency (and therefore responsibility) of the key individuals who brought the trials about.[xxiii] Other historians, particularly those who prefer psychological or pathological explanations such as hysteria or conversion disorder are more likely to describe the events of 1692 as an ‘outbreak’.[xxiv] The key point here is that it simply does not make sense to strictly delineate historical ‘description’, ‘explanation’, and ‘argument’. How one chooses to describe historical phenomena constitutes part of one’s historical causal explanation and it all needs to be argued.

So we appear to have a situation where some history examination questions fundamentally misrepresent historical argument as written by academic historians.

What exacerbates the issue, however, is when examination genres that do not treat academic history as the gold standard are extrapolated to form genre-based progression models. For example, we find genres used in this way in progression guidance endorsed by exam boards[xxv] or elaborate ‘genre-based pedagogies’ where these examination genres are used to justify linguistically-informed ‘pedagogical pathways’.[xxvi] Such progression models usually take a form such as;


Screenshot 2018-10-25 19.01.24[xxvii]

Such models often advocate Key Stage 3 students spend much of their compulsory history education writing non-historical, non-argumentative ‘descriptions’ or ‘categorical explanations’ before they ‘progress’ to written argument.

Of course, it is unreasonable to assume teenagers will write to the standard of academic historians. ‘Academic’ and ‘school’ history necessarily cannot be the exact same species. But while these models have a superficially seductive linearity I’ve yet to encounter a persuasive argument as to why it follows that drilling students to write in non-historical genres is necessary or even unharmful in enabling students to write sophisticated historical arguments; especially as we have evidence that primary students can appreciate aspects of history’s argumentative nature.[xxviii]

Perhaps more crucially, what are the essential ramifications if older students are examined in non-historical genres and those genres form the basis of progression models for younger students? A danger is that ‘school history’ ceases to be a simplified, ‘recontextualised’ version of academic historical knowledge – accessible to younger students but of essentially the same genus as what is done by academic historians – but instead becomes an object of study of a totally different order.


[i] Stone, L. (1987). The Past and the Present Revisited. London: Routledge & Kegan Paul.p.75

[ii] Atkinson, R. F. (1978). Knowledge and Explanation in History: An Introduction to the Philosophy of History. London: The MacMillan Press p.136

[iii] Pearson (2017). Advanced Paper 3: Themes in breadth with aspects in depth Option 33: The Witchcraze in Britain, Europe and North America c1580–c1750

[iv] Pearson (2014) Level 3 Advanced GCE in History (9HI0) Sample Assessment Materials 

[v] Marius, R. & Page, M. E. (2007). A Short Guide to Writing About History Pearson Longman: London pp. 56-57

[vi] Tosh, J. (2006). The Pursuit of History: Aims, methods and new directions in the study of modern history London: Pearson p.156

[vii] Byrom, J. (1998). ‘Working with Sources: Scepticism or Cynicism? Putting the Story Back Together Again’. Teaching History 91 pp.32-35; Lang, S. (2003). ‘Narrative: the under-rated skill’ Teaching History 110 pp.8-15. For an introduction to these debates try Counsell, C. (2016). ‘New, Novice, or Nervous: Constructing Narrative’ Teaching History 164 p.57

[viii]  Department for Education (2013). History programmes of study: key stage 3 National curriculum in England p.1

[ix] Pearson (2015). Sample Assessment Materials (Question Papers) Pearson Edexcel Level 1/Level 2 GCSE (9-1) in History (1HI0) pp. 4 & 43-44

[x] For example, it is used for ‘Explain What’ with Similarity and Difference (p.7); Evidential Thinking (p.19); and Change and Continuity (p.24). It is also used for ‘Explain Why’ for Causation (p.8). Ibid.

[xi] Ibid. pp.43-44 & 49-50

[xii] Ibid.

[xiii] Goldstein, L. J. (1976). Historical Knowing London: University of Texas Press. pp. 142-143

[xiv] Megill, A. (2007). Historical Knowledge, Historical Error: A Contemporary Guide to Practice. The University of Chicago Press: London pp.87-88

[xv] For a discussion of history as argument try Megill’s ‘unresolving dialectic’. Megill, op. cit. p.2

[xvi] Fischer, D. H. (1970). Historians’ Fallacies: Towards a Logic of Historical Thought. New York: Harper Perennial.  p.vx

[xvii] Pearson (2015) op. cit. p.4

[xviii] Elton, G. R. (1989). The Practice of History. London: Fontana Press. pp.175-176

[xvix] Hewitson M (2015). History and Causality. London: Palgrave Macmillan.

[xx] Walsh, W. H. (1958). ‘”Plain” and “Significant” Narrative in History’ The Journal of Philosophy of History 55 pp.479-484

[xxi] Megill op. cit. pp.87-88.

[xxii] Norton, M. B. (2002). In the Devil’s Snare: The Salem Witchcraft Crisis of 1692. New York: Alfred A. Knopf. p.8; E.W. Baker A Storm of Witchcraft: The Salem Trials and the American Experience (2015) pp.12-13.

[xxiii] Fels, T. (2017). Switching Sides: How a Generation of Historians Lost Sympathy for the Victims of the Salem Witch Hunt. Baltimore: John Hopkins University Press. p.128

[xxiv] For example Baker op. cit.pp.12-13 & 99-102; Hansen, C. (1988). Witchcraft at Salem. London: Arrow Books. pp.145-146

[xxv] Pearson Key Stage 3 Exploring History: Inspire young historians and build the foundations for Pearson Edexcel GCSE (9-1) History

[xxvi] I discuss how generic ‘genre-based’ pedagogies tend to use exam boards as the foundations for their ‘pedagogical pathways’ here Carroll, J. E. (2018) ‘Epistemic explanations for divergent evolution in discourses regarding students’ extended historical writing in England’, Journal of Curriculum Studies, DOI: 10.1080/00220272.2018.1499805

[xxvii] For example Coffin, C. (2006). Historical discourse: The language of time, cause and causation. London: Continuum; Christie, F., & Derewianka, B. (2008). School discourse. London: Continuum.

[xxviii] I summarise some examples of students aged 9-13 arguing historically in Carroll op. cit. p.10

Setting the Story Straight Part 1: In Defence of Narrative for some A-Level Cause Questions

In each of the last two years, the A-Level Examiners’ Report for my students’ Witch Craze module has recommended that ‘candidates should avoid a narrative/descriptive approach; this undermines the analysis that is required for higher levels’.[i] Despite the examiners’ retroactive recommendations, neither the mark scheme nor the wording of the questions specify that students should argue in the analytic mode.[ii]

In this first part of a two-part post I want to argue that, in my view, there’s one essential problem with this overgeneralising statement if we use academic history as our ‘gold standard’ of historical explanation. The guidance actively discourages students from writing in the mode that many historians write in for certain types of causal questions. As Marius and Page suggested, ‘without narratives, history would die as a discipline’.[iii] A student who has been drilled to believe that narrative/description ≠ causal explanation might be confused to find the majority of causal explanations in the historiography of the Salem witch trials are argued, at least partially, in that very mode. The student’s likely confusion might increase further when we consider that the academic texts that teachers usually recommended to students at A-Level tend to be those we consider to be ‘engaging’ and ‘accessible’ for the general reader – in other words, narratives.[iv]

The exam board’s suggestion applied, in part, to essays covering ‘in-depth’ case studies which could involve causal questions covering a time period of less than a year. For example, one of the 2017 questions in this section was;

How far does the role of Governor Phips explain the end of the Salem witch hunt (1692–93)? [v] 

For now, it is important to note that this question demands the student spends part of their essay explaining how one individual (Governor William Phips) contributed to the ending of the Salem witch trials over a relatively short time span of approximately nine months from September 1692 to the trials’ end in May 1693.

A student answering this question in the analytic mode might, depending on what their argument regarding the most important cause was, structure their response rhetorically as follows;

Screenshot 2018-10-25 19.01.24Note that this approach turns an individual’s actions that point toward an explanation of the end of the trials (e.g. Phips closed the Court of Oyer and Terminer in October 1692) into a broad configurational abstraction that doesn’t actually explain very much (‘Phips’ role’). One way around this would be to create better abstractions that are explanatory (e.g. ‘Phips’ crucial role in ending the legal process’). Another way, probably in conjunction with a better abstraction, could be to focus on Phips’ actions in detail.

Note too that this approach necessarily hinders explanation of why Phips was motivated to act the way he did by downplaying the importance of chronology. For example, some historians suggest one reason why Phips closed the Court of Oyer and Terminer was because in the summer of 1692 the afflicted girls began ‘overreaching’ and calling out higher-status individuals including Phips’ own wife.[vi] Although it is perfectly reasonable to argue that the girls’ actions were less important than Phips’ in bringing the trials to an end, we can only truly explain what motivated Phips by detailing what happened before he took his actions. This ‘analytic’ structure, at the very least, obscures that opportunity to explain by ignoring the sequence of his decision-making.

Despite the exam board’s categorical claim that the narrative/descriptive mode ‘undermines’ explanation, for this particular type of question there is far from a consensus among theorists of history that the analytic mode would be the best to adopt. In fact, Roberts argued that this type of discussion represents ‘the most important and central debate in the philosophy of history since the 1960s: the extent to which the discipline of history is essentially a narrative mode of knowing, understanding, explaining and reconstructing the past’.[vii]

The exam board’s stance does mirror intense criticism of the narrative mode in the mid-twentieth century which constituted, according to Ricouer, the ‘eclipse of narrative’.[viii] Methodologically, narrative was attacked by the likes of Annalistes, Freudian psycho-historians, and clioemtricians for the mode’s inability to explain in a specifically ‘scientific’ sense. [ix] Ontologically, these analytic historians tended to privilege slower-moving, longer-term structural conditions in their explanations such as economies, mentalités, and geography which, it was argued, were unsatisfactorily accommodated by narrative.[x] It was during this period that the ‘prevalent suspicion that narrative as such is epistemologically defective’ emerged. [xi] We can see, I think, vestiges of this viewpoint in the exam board’s guidance.

Narrative’s ‘eclipse’ was mirrored in the historiography of Salem in the 1970s and 1980s when the psychological, social, and economic dimensions of the witchcraft crisis were emphasised and historians employed a number of social scientific techniques from sociology, anthropology, and psychology.[xii]

Towards the end of the twentieth century, however, a counter-attack by narrative historians was launched which Stone dubbed ‘the revival of narrative’.[xiii] In its strongest form, some narratavists went so far as to argue that narrative was the only mode that facilitated a specifically historical explanation.[xiv] The defence of narrative has been variegated and extensive but generally narratavists have tended to be sceptical of the imposition of a positivist-scientific model of explanation onto the discipline of history.[xv]

So why might the narrative/descriptive mode be particularly appropriate for explaining the actions of individuals over short time-spans? This view has its roots particularly in Oakeshott’s ‘continuous series model’ of explanation with its dictum that ‘history accounts for change by means of a full account of change’.[xvi] In this strong view, far from ‘undermining’ historical explanation, the narrative/descriptive mode is essential in explaining human actions in the past.  From this perspective, historians usually can’t rely on deduction from universal, impersonal, atemporal laws to explain temporally-bound, unique, human actions.

Consider how an experimental scientist and a historian might seek to ‘explain’ the same event. A person falls from a window. The physicist will probably explain this event with a universal, impersonal, abstract law – ‘gravity’. The explanation is atemporal in the sense that it doesn’t matter if the person fell in 1618 in Prague or yesterday in London. In this sense, it’s a generalising whenever explanation rather than an individualising when explanation.

A historian, however, will probably take ‘gravity’ for granted and instead want to explain this specific person’s action at this specific point in time. Did they jump? If so, why? It’s crucial we ascertain what happened before to explain what happened after. At this point, some descriptive statements alone such as ‘Two days ago they told their best friend they were having suicidal thoughts due to money problems. Last night someone saw them climb onto the ledge of the 10th floor. They fell.’ might be enough to satisfy most people’s (including most readers of history’s) need for an explanation. In this particular case focusing on short-term individual actions, chronological description still explains. We could pepper the description with ‘causal connectives’ such as ‘so’, ‘therefore’, and ‘because’ but some theorists of history at least consider that to be redundant and inelegant. [xvii]

Let’s return to our example of Phips ending the Salem witch trials.  According to Oakeshott, instead of covering laws to bridge the gap between cause and consequence we have to rely on description of other intervening events. While doing so, the causes become clearer because the events are selected by virtue of their relevance to the question and their interrelationship with one another. So, for example, we might quickly plan a narrative response by filling in more detail about Phips’ key actions in ending the witch hunt.

Screenshot 2018-10-25 19.01.32

This plan, however, still doesn’t explain that much. Why, after all, did Phips dissolve the Court of Oyer and Terminer that he himself had created in May 1692? And why did he suddenly turn against the use of spectral evidence? More chronological description here helps point toward an explanation.

Screenshot 2018-10-25 19.21.37

But why might Phips have been motivated to act based on Increase Mather’s advice?

Screenshot 2018-10-25 19.28.13

In this strong view, the describing ‘how’ is inseparable from the explaining ‘why’ if a historical description is so complete that it permits no lacuna. Clearly, this ‘continuity’ in description is more achievable in shorter rather than longer timescales. This mode, according to Oakeshott, ultimately renders overt ‘explanation’ in the analytic mode superfluous.[xviii]

According to its advocates then narrative is key to explaining individual actions and events.[xix] For example, the mode enables the historian to empathise with the agents they are studying which in turn helps explain their actions. [xx] This empathy is achievable, according to narrative’s advocates, because this mode better approximates the sense of time we experience in our own lives so we can better appreciate the circumstances in which the agents made their choices. [xxi] This claim often reflects a broader belief that structural analyses borrowed from the social sciences encourage static rather than dynamic explanations that downplay ‘time’s arrow’ which is so vital to history.[xxii]

Accordingly, the historiography of Salem witnessed a spate of explicitly narrative histories around the crisis’ tricentennial anniversary in the 1990s such as those by Gragg, Rosenthal, Hill, and Hoffer.[xxiii] For example, Le Beau noted his work;

employs a narrative format that was once popular among historians but that has fallen out of favour. Such an approach is intended to make the subject more accessible to the reader, to recapture some of the drama that has held people spellbound for so long, and to suggest yet another way of looking at the event. As Larry Gragg has reminded us, to fully appreciate what happened in 1692, we must ‘explore the particular decisions made by individuals involved and their consequences’. When all other avenues of interpretation have been exhausted, we are left with the fact that individuals and individual decisions matter.[xxiv]

One wonders how well Le Beau would have done in the 2017 A-Level exams with such an approach.

Settling the debate as to what the best mode of historical explanation is far beyond my powers. I suspect, however, that the lack of a universally agreed explanatory model in history means that it is ultimately irresolvable. My only goal at this juncture is to try and point out that, despite the exam board’s prescriptions, exclusivity-claims regarding the ‘best’ mode for historical explanations are, at the very least, debatable if we take academic history as our benchmark for what constitutes ‘good’ historical explanation. It is therefore not clear why the mode should be discouraged as such if, a) when done well, it can be highly explanatory and b) there is no mention that the narrative mode will be penalised in the mark scheme.

In my next post, I will outline some possible ramifications for how we might teach and assess extended writing in history if one is willing to accept this initial premise.

[i]Pearson (2017) Examiners; Report June 2017 GCE History 9H10 33 p.35;; Pearson (2018) Examiners; Report June 2018 GCE History 9H10 33 p.41;

[ii] Pearson (2017) Mark Scheme Summer 2017 Edexcel in History (9H10/33) Advanced Unit 1: Themes in Breadth with Aspects in Depth Option 33: The Witchcraze in Britain, Europe and North America c.1580-c.1750

[iii] Marius, R. & Page, M. E. (2007). A Short Guide to Writing About History Pearson Longman: London p.61

[iv] A number of historians have noted that narrative histories tend to appeal more to the general reader, as evidenced by the fact that narrative histories tend to sell far better than their analytic counterparts. Black. J. & MacRaild (1997). Studying History London: MacMillan; Stanford, M. (1998). An Introduction to the Philosophy of History. Oxford: Blackwell. Historians have also noted that narrative tends to be more accessible because it usually focuses more on the structure of individuals’ actions over shorter time scales so it is more congruent with how the reader experiences their own lives. Roberts, G. (1996). ‘Narrative History as a Way of Life’ Contemporary History 31(1) 221-228; Stanford, (1998) op. cit.; Tamura, E. H. (2011). ‘Narrative History and Theory’. History of Education Quarterly, 51(2) 150-157

[v] Pearson (2017). Advanced Paper 3: Themes in breadth with aspects in depth Option 33: The Witchcraze in Britain, Europe and North America c1580–c1750

[vi] Hill, F. (1996). A Delusion of Satan: The Full Story of the Salem Witch Trials (1996).

[vii] Roberts, G. (2001) (Ed.). The History and Narrative Reader London: Routledge. p.1

[viii] Ricouer, P. (1984). Time and Narrative Vol. 1 Chicago IL: University of Chicago Press. p.95ff

[ix]Barzun, J. & Graff, H. F. (1985). The Modern Researcher 4th Edition. London: Harcourt Brace Jovanovich; Stone, L. (1987). The Past and the Present Revisited. London: Routledge & Kegan Paul.

[x] Carr, D. (2008). ‘Narrative Explanation and its Malcontents’ History and Theory 47 pp.19-30; Black & MacRaild op. cit. 

[xi] Megill, A. (2007). Historical Knowledge, Historical Error: A Contemporary Guide to Practice. The University of Chicago Press: London. pp.87-88

[xii] Gragg, L. (1992). The Salem Witch Crisis London: Prager.

[xiii] Stone, L. (1979). ‘The Revival of Narrative: Reflections on a New Old History; Past & Present 85 3-24.

[xiv] Gallie, W. B. (1964). Philosophy and the Historical Understanding. London: Chatto and Windus

[xv] Roth, P. A. (1988). ‘Narrative Explanations: The Case of History’ History and Theory 27(1) pp.1-13

[xvi] Oakeshott, M. (1933). Experience and its Modes Cambridge: Cambridge University Press p.143

[xvii] For example, Davidson suggests that in such circumstances ‘an apparently simple string of details may convey knowledge more elegantly, more directly, and with more impact than a general statement of any sort’. Davidson, J. W. (1984). ‘The New Narrative History: How New? How Narrative?’ Reviews in American History 12(3) pp.327-328

[xviii] Oakeshott, op. cit. p.143

[xix] For example Vincent (2005) argued ‘structure is not enough; by omitting narrative, one omits truth, by omitting or understanding the part played by chance, uncertainties, events, contingencies. Structure may set limits to what can happen…but it can never by itself explain why one course of events took place rather than another’ (p.89).

[xx] Stone (1987) op. cit.

[xxi] Tosh, J. (2006). The Pursuit of History: Aims, methods and new directions in the study of modern history London: Pearson

[xxii] Munz, P. (2006). The Historical Narrative. In Bentley, M. (Ed.) Companion to Historiography (pp.851-872). London: Routledge; Stanford, M. (1994). A Companion to the Study of History. Oxford: Blackwell.

[xxiii] For example, Demos recognised this re-emergence by noting that the tricentennial anniversary of the crisis coincided with a surge of ‘narrative retellings of the trial sequence’ Demos, J. (2008). The Enemy Within: 2,000 Years of Witch-Hunting in the Western World. Viking: London. (pp.207-208).  Examples of the narrative histories include Gragg op. cit; Rosenthal, B. (1993). Salem Story: Reading the Witch Trials of 1692. Cambridge: Cambridge University Press.; Hill op. cit.; Hoffer, P. C. (1996). The Devil’s Disciples: Makers of the Salem Witch Trials. London: John Hopkins University Press; and  Hoffer, P. C. (1997). The Salem Witchcraft Trials: A Legal History. Lawrence, Kansas: University of Kansas Press.

[xxiv] Le Beau, B. F. (1998) The Story of the Salem Witch Trials Prentice Hall: Upper Saddle River, New Jersey. pp.ix-x

Duplo to watercolours: how the substantive might shape the disciplinary in students’ historical causal arguments

In a previous post, I argued that writing frames might be more effective when the teacher is clear regarding the particular historical second-order concept they want their students to focus on. For example, it might be beneficial to be precise about whether one is concentrating on the specific language of historical causation rather than more generic ‘explaining’ language associated more broadly with ‘making clear’, ‘explication’, or ‘elucidation’.  While it is clearly useful for students to be comfortable using connectives such as ‘also’ and ‘however’, these frames recommended by centralised initiatives such as the National Strategies for Literacy in History might only get students so far in constructing historical arguments distinct from those in science, geography, or indeed any other ‘non-fiction genre’.[i]

In writing that post, I was very deliberately trying to operate in a rich tradition of previous history-teacher curricular theorisation. As I discussed here, a number of history teachers in England from the mid-2000s became (either consciously or unconsciously) dissatisfied with generic writing frames that were falsely advertised as the language of historical argument.[ii] Instead, a number of these teachers adopted history’s second-order concepts as an investigative framework for their theorisation regarding language instruction. Teachers such as Woodcock (causation);[iii] Foster (change and continuity);[iv] Bradshaw (similarity and difference);[v] and many others began to identify the specific language students might require in order to respond in writing to the type of questions historians, by disciplinary convention, tend to argue about.

My current research involves trying to identify the language of causation that historians use in their arguments so that I might explicitly teach it to my students. While thus far I have been specifically focusing on the historiography of one particular topic (the Salem witch trials of 1692), I have found certain ‘chunks’ of causal language that I think might be generally applicable in causal arguments across history generally (see below). Similarly, I have also identified language regarding characterisation of causes (e.g. ‘precondition’[vi]); explanation of multicausality (e.g. ‘a perfect storm of factors’[vii]); or prioritisation of causes (e.g. ‘most of important of all’[viii]) that might also be transferable in this way.

Word grid causes

Any of this kind of subject-wide yet conceptually-precise language that historians use is, in my view, extremely useful for students to learn. But it is important to remember another reason why history teachers such as those I mentioned above began to move away from resources such as the National Strategies for Literacy.  Writing frames shorn from the specific context of a legitimate historical enquiry can lead to a marginalisation of the shaping effect particular substantive topics might have on historical arguments.[ix] Just as the teacher needs to consider the distinction between the generic and the historical, similarly they might need to consider the particulars of the higher-resolution of historical topics. For instance, the uncritical use of language mats – even if they are specific to a particular second-order concept such as causation – might lead to the teacher inadvertently deemphasising disciplinary language specifically valuable to causal enquiries on particular topics such as the Salem witch craze.

Let’s take a hypothetical example. Citing instances such as Elizabeth Knapp in Groton in 1672; the Goodwin children in Boston in 1688; and Kate Branch in Stamford in 1692; many historians of Salem have noted that while children and young adults suffering afflictions and making accusations of witchcraft in 17th century New England was unusual, it certainly was not unheard of either. Yet none of these other cases escalated into a full-blown witch hunt. One cause that distinguished Salem in 1692 was that powerful male figures such as the girls’ guardians (e.g. Samuel Parris and Thomas Putnam); examining magistrates (e.g. John Hathorne and Jonathan Corwin); and presiding judges (e.g. William Stoughton) were willing to give the Salem girls’ accusations an unusually high amount of credibility.

In explaining the importance of these men’s role, a student relying solely on history-wide causal language such as the chunks I displayed above might write something like this;

Puritan leaders were worried that people in New England were becoming less devout which led to powerful male leaders in Massachusetts looking for scapegoats. Consequently, these leaders’ anxiety meant they were willing to listen to the girls’ accusations which caused the Salem witch hunt.

This example is, I think, fine. But how might we improve it? In my view, it has a certain artificiality because it is a manifestation of a type of ‘Duplo’ approach to constructing historical argument. First, take a substantive ‘fact’ block. Then slot in a general disciplinary ‘causal language’ block next to it. Repeat. Hey presto! We have a causal argument (see below).


Recently, a number of history teachers have begun to consider how a strong command of substantive knowledge manifests itself in students’ historical arguments. For example, in an already-seminal article Hammond discussed how strong substantive knowledge ‘flavoured’ her students’ historical claims.[x] In this vein, I’ve been looking at the historiography of Salem to see possible examples of how historians’ substantive knowledge ‘flavours’ their choice of disciplinary language so that I might be able to model it for my students. Consider the following example from Enders A. Robinson;

At the strategic level, the old-guard Puritans granted the authority under which the conspiracies operated. The beginning of 1692 saw the old guard running an outlaw government in New England, these men were the councillors, magistrates, judges, and high military officers. Only the old guard had sufficient authority to sanction the atrocities of the witch hunt.[xi]

Note Robinson’s choice of ‘granted authority’ and ‘sanctioned’ instead of, for instance, ‘resulted in’ or ‘led to’. We can find many other historians using similar verbs and nouns when explaining the Puritan old-guard’s role in the Salem witch hunt: ‘gave credence’;[xii] ‘gave (their) seal of approval’;[xiii] ‘affirmed’;[xiv] ‘endorsement’;[xv] ‘gave (their) solid backing’;[xvi] ‘lent their weight to’;[xvii] ‘gatekeeper’;[xviii] ‘put their weight behind’;[xix] ‘advocate’;[xx] or ‘gave the green light’.[xxi] In this particular context, this language is all clearly causal despite the fact that such nouns and verbs might not necessarily make sense in causal arguments in other historical topics. This lack of transferability is because this causal language is particularly ‘flavoured’ by substantive knowledge of the societal power structures of the late-17th century Puritan New England patriarchy. The girls’ accusations could not have led anywhere unless these powerful men had validated them. That potentially ‘flavoured’ argument, however, is shut down by the exclusive use of ‘flavourless’ language such as ‘due to’, ‘meant’, or ‘brought about’.

(Sidenote. The fact that historians often ‘hide the analytical ductwork’[xxii] by opting to use topic-specific causal verbs and nouns rather than overt and general ‘analytical’ language such as ‘led to’ might be one reason why academic history can often read as if it is narrative/descriptive while at the same time being highly explanatory and analytical).

So unlike the ‘Duplo’ model authentic historical writing is often more like working with two watercolour paints and sometimes failing to stay within the lines. To varying degrees, the colours often bleed into one another (see below);


In terms of planning, this enquiry specificity might have at least two ramifications for history teachers. First, a teacher can’t necessarily assume the language mats or word grids one might see on social media, in examination board guidance, or teacher-resource marketplaces can be treated as ‘off-the-peg’.[xxiii] The ones I have designed, for instance, were very deliberately created with particular enquiries in mind.[xxiv] But what may have worked effectively for my students arguing why the Nazis rose to power will not automatically be suitable for another teacher’s students explaining why William won the Battle of Hastings. The substantive content of one might shape the disciplinary in ways that would be inappropriate for the other. Instead, the teacher must use their professional judgement to use, adapt, or discard as required.

Ultimately, no language mat in itself can ever be a panacea. In fact, as previous history teachers have persuasively demonstrated grids might only be of limited use unless situated in the context of an enquiry that makes clear the disciplinary and substantive knowledge that sits behind such language’s use. Like any proxy, unless the rationale for its construction is understood any potential effectiveness will almost certainly be lost in transit.

Furthermore, teachers might also consider highlighting a small number of specific verbs or nouns especially relevant to the causes students will encounter in that enquiry that historians have used in their interpretations. Such language might make students move away from ‘Duplo’ writing toward authentic academic argument. It could also highlight to students the subtle interpretative choices historians make. But it might also serve to reinforce the substantive knowledge the students are required to know. Playing around with deceptively simple language such as ‘gave the green light to’, for example, might provide another access point to students struggling to grasp the notion of the Puritan old-guard’s gatekeeping role in this type of society.  In this sense, might not the linguistic also release the substantive?

A final thought. Language mats run the risk of encouraging students to visualise conceptual thinking as a ‘process’ or ‘skill’.[xxv] It might be all too easy for a student to think that when they encounter a new topic all they need do to answer a causal question is reach for their trusty toolbox, grab the appropriate ‘tool’ (i.e. the language mat), and apply the same repeatable, tried-and-tested processes for explaining why things happened in history. Such approaches, however, might lead the student to believe that historical substantive content is free-floating, inert information that does not affect the disciplinary; essentially another interchangeable slab students just need to set to work on. But this is not how historical argument always works. In reality, the disciplinary and the substantive often do not operate in discrete silos. Disciplinary conventions and unique substantive content meet, mesh, and modify one another in in order to construct historical knowledge.

[i] Department for Education and Skills. (2002). Key stage 3 national strategy – literacy in history. London: DfEaS & 21 Retrieved from

[ii] Carroll, J. E. (2017). ‘From divergent evolution to witting cross-fertilisation: the need for more awareness of potential inter-discursive communication regarding students’ extended historical writing’ The Curriculum Journal 28(4) pp.504-523

[iii] Woodcock, J. (2005). ‘Does the linguistic release the conceptual? Helping Year 10 to improve their causal reasoning’ Teaching History 119, pp. 5-14.

[iv] Foster, R. (2013). ‘The more things change, the more they stay the same: Developing students’ thinking about change and continuity’. Teaching History 151, pp. 8–16.

[v] Bradshaw, M. (2009). ‘Drilling down: How one history department is working towards progression in pupils’ thinking about diversity across Years 7, 8 and 9’. Teaching History, 135, pp. 4–12.

[vi] Baker, E. W. (2015). A Storm of Witchcraft: The Salem Trials and the American Experience. Oxford: Oxford University Press pp.123-125

[vii] Ray, B. C. (2015). Satan and Salem: The Witch-Hunt Crisis of 1692. London: University of Virginia Press p.5

[viii] Hansen, C. (1988). Witchcraft at Salem. London: Arrow Books. pp.145-146

[ix]Counsell, C. (2011a). History teachers as curriculum makers: professional problem-solving in secondary school history in England. In Schüllerqvist, B. (Ed.) Patterns of Research in Civics, History, Geography and Religious Education (pp.53-88). Kalrstad: Kalrstad University Press.

[x] Hammond, K. (2014). ‘The knowledge that ‘flavours’ a claim: Towards building and assessing historical knowledge on three scales’. Teaching History 157: Assessment, pp. 18–25.

[xi] Robinson, E. A. (1991). The Devil Discovered: Salem Witchcraft 1692. New York: Hippocrene Books.p.252

[xii] Hill, F (1996). A Delusion of Satan: The Full Story of the Salem Witch Trials. London: Penguin p.69

[xiii] Ibid

[xiv] Rosenthal, B. (1993). Salem Story: Reading the Witch Trials of 1692. Cambridge: Cambridge University Press pp.6-7

[xv] Ray op. cit. pp.46-47

[xvi] Ibid.

[xvii] Norton, M. B. (2002). In the Devil’s Snare: The Salem Witchcraft Crisis of 1692. New York: Alfred A. Knopf pp.306-308

[xviii] Ibid. p.72

[xix] Hoffer, P. C. (1996). The Devil’s Disciples: Makers of the Salem witchcraft trials. John Hopkins University Press: London.  p.103

[xx] Ray op. cit. pp.144-145

[xxi] Baker op. cit. p.30

[xxii] Lee, P. & Shemilt, D. (2009). ‘Is any explanation better than none? Over-determined narratives, senseless agencies and one-way streets in students’ learning about cause and consequence in history’ Teaching History 137, 42-49.

[xxiii] For example, one exam board has – without my permission, without crediting me, and without linking the article from which the resource came from – disseminated one of my resources suggesting that my language mat is applicable across topics when answering GCSE questions in the narrative mode. The resource was, in fact, not designed for this purpose at all. It was designed for causal A-Level arguments in the explanatory mode for students studying the Nazi rise to power. The resource, therefore, will likely be only of limited use for what the exam board claims.

[xxiv]  For example Carroll, J.E. (2016) ‘The whole point of the thing: how nominalisation might develop students’ written causal arguments’ Teaching History 162 pp. 16-24

[xxv] Both Counsell and Lee have discussed dangers of this tendency e.g. Counsell, C. (2000) Historical knowledge and historical skill: the distracting dichotomy. In Issues in History Teaching ed. J. Arthur and R. Phillips, 54-71. Abingdon: Routledge; Counsell (2011) op.cit.; Lee, P. (2005) Putting principles into practice: understanding history. In How Pupils Learn: History in the classroom, ed. M. S. Donovan and J. D. Bransford, 31-78. Washington DC: National Academies Press; Lee, P. (2011) History education and historical literacy. In Debates in history teaching, ed. I. Davies, 63-72. London: Routledge.








‘Another reason why…’ some writing frames stunt students’ historical causal arguments


In a series of previous posts, I have argued that history teachers portraying sentence starters such as ‘I believe…’ and ‘explaining’ connectives such as ‘so’, ‘therefore’ or ‘because’ as ‘the language of the historian’ is misleading. I tried to demonstrate how, far from being ‘how historians write’, such frames at best limit students’ scope to construct their own historical arguments. At worst, some might be fundamentally anti-historical.

Continuing this theme, in this post I want to consider ‘paragraph starters’ such as;

‘One reason why…’

‘Another reason why…’

‘A final reason why…’[i]

By including ‘reason why’ these starters seem most appropriate for arguments constructed in response to questions focused on the second-order concept of causation. Presumably, they could be tweaked to be applicable to other conceptual focuses; perhaps ‘Another reason why Cromwell was significant was…’ for significance or ‘A final difference between the nobility and the gentry was…’ for similarity and difference.  For now, however, let us assume that in these frames ‘reason’ operates synonymously with ‘cause’. An example of a causal question that the frames might be used in response to could be;

How far does the role of Governor Phips explain the end of the Salem witch hunt (1692–93)? [ii]

In this question, the student would first be expected to recognise the cause given and argue its relative importance. For example, they might consider how upon returning from fighting the Wabanakis on the frontier in September 1692 Phips closed the Court of Oyer and Terminer which had admitted spectral evidence contributing to a 100% conviction rate. Phips then established the new Superior Court of Judicature in its place which did not admit this type of evidence and therefore predominantly acquitted.

The students would then be required to argue whether other causes which contributed to the craze subsiding were more or less important. Although it is an interpretation which has been rather debunked by more recent research, a traditional cause given for the hunt coming to an end was the ‘afflicted’ girls’ ‘overreaching’: their accusing of those in positions of power including, perhaps, Phips’ own wife. Fearing that the girls might now turn their ferocity on them, powerful male societal gatekeepers removed their endorsement of the girls’ claims. In this staunchly patriarchal society, the girls’ accusations ceased to carry weight once shorn of this legitimation.

A further cause the students might include in their explanatory arguments could be the increasingly loud and convincing campaign led by more moderate and influential Puritan ministers – most famously Increase Mather. From the comparative safety of Boston, men such as Mather began to publicise arguments against the admission of spectral evidence helping to convince Phips, who was a member of Mather’s congregation, to change tack.

For our purposes here, let us limit ourselves to these three causes of the end of the hunt. Teachers might instruct students to adopt the following essay structure;


One reason why the Salem witch hunt ended in 1693 was…the intervention of Governor Phips…

Another reason why the Salem witch hunt ended in 1693 was…the ‘afflicted’ girls’ ‘overreaching’…

A final reason why the Salem witch hunt ended in 1693 was… influential figures in Boston criticising the admission of spectral evidence…


These frames are an example, I think, of how Australian genre theory has influenced literacy instruction in English secondary schools – albeit often in a bowdlerised, indirect, and uncredited way.[iii] At the very least, these scaffolds seem to have evolved in parallel to some genre theorists’ strikingly similar recommendations.  Such suggestions include students starting paragraphs in ‘categorical explanations’ with ‘numeratives’ and a clear ‘hyper-theme’ such as ‘A second reason was (insert cause here)’.[iv] The rationale for such modelling is that it highlights for students the engendering or creation of cohesion and structure in texts.

Encouraging students to structure their extended writing coherently is clearly a laudable aim and I am not suggesting that some scaffolding, used judiciously, should not be used to this end. I do think, however, that there is an issue with these particular frames which seem at best only generically applicable to explanation in all ‘non-fiction’ writing. They present historical explanation as of essentially the same order as explanations in other disciplines where there might not be the same disciplinary conventions demanding one argues the relative importance of causes. Such frames, for example, might be perfectly appropriate for reporting why fires burn: ‘One reason is a source of heat…’. ‘Another reason is fuel…’.  ‘A final reason is the presence oxygen…’. What matters in the fire example is that all of these causes are necessary in combination, not whether one is more important than the other.

But in a historical causal explanation a lack of attentiveness to the specifics of the discipline, reified in frames such as ‘another reason why was the girls’ overreaching…’, will almost always have a deadening effect. What most historians care about are problems such as whether the girls’ overreaching was more important than Phips’ return or not.

It is accepted by many historians that the main driver of a causal argument is the prioritisation of causes in terms of relative significance into some form of ‘pecking order’[v] or ‘hierarchy’[vi]. In doing so, these historians suggest one should aim to argue what the most important cause was. Carr perhaps most forcibly encapsulated this view regarding the typifying characteristic of historical causation;

The true historian, confronted with this list of causes of his own compiling, would feel a professional compulsion to reduce it to order, to establish some hierarchy of causes, which would fix their relation to one another, perhaps to decide which cause, or which category of causes, should be regarded ‘in the last resort’ or ‘in the final analysis’ (favourite phrases of historians) as the ultimate cause, the cause of all causes. This is his interpretation of his theme; the historian is known by the causes which he invokes…every historical argument revolves round the question of priority of causes.’[vii]

According to historians such as Hewitson, without such ‘ordering of ‘causes’ according to their salience or significance in respect to the question posed’ causal argument ‘itself would not be possible’. [viii]

If we accept this then while ‘another reason’ provides structural coherence it is completely stultifying to causal historical argument. In this sense I see starters such as ‘another reason why’ as a stark manifestation of a dangerous mentality that tends to view ‘the literacy’ as ‘a skill’ somehow disconnected from ‘the history’.[ix] In this instance, the desire to emphasise the importance of structure (‘the literacy’) might disable students from arguing the relative importance of causes (‘the history’). In particular, the frames might blinker the student from the fact that some historians a) use explicit language to argue the relative importance of causes and b) might rhetorically organise their overall text in such a way as to perpetuate their overall causal arguments.

Let us return to our example of the Salem witch craze’s loss of impetus. Historians in this historiography have often displayed the ‘professional compulsion’ described by Carr to prioritise their causes. For example, after outlining and explaining the role of less important causes Chadwick Hansen emphasised that, in his view, the girls’ ‘overreaching’ was the most important cause in the craze ultimately fizzling out;

But most important of all, as the witch hunt spread and the accusations flew, people were accused whom nobody could think guilty.[x]

Here, Hansen emphasised his argument by building toward the most important cause and indicating it clearly. A historian might structure their overall writing this way, perhaps beginning by arguing the limitations of existing arguments and thus preparing the reader for the clear statement and elaboration of the historian’s main thesis regarding the most important cause.

Frances Hill instead opted to emphasise what, from her perspective, was most crucial in ending the craze (the campaign in Boston against the Court of Oyer and Terminer’s admission of spectral evidence led by Increase Mather) by leading with the most important cause.

A huge backlash ended the witch-hunt. The single most important element in that backlash was an essay composed by Cotton Mather’s father, Increase called Cases of Conscience Concerning Evil Spirits Personating Men.[xi]

As both these examples show when arguing the relative importance of causes historians in the historiography of Salem might use the organisation of their text as an argumentative tool. Some also use clear language to indicate their overall argument in this regard. Both of these possible articulations of argument are shut down by prescribed essay structures based on starters such as ‘another reason why…’ which either imply the causes are of equal importance or even worse still might suggest to the student that prioritisation is not required at all.

I have been trying to identify the specific language historians of Salem such as Hansen and Hill use in their causal arguments. I select particularly argumentative extended extracts to read with my class and then I model the authentic language the historian used. Below you can see some of the ‘chunks’ I have identified so far.[xii] Chunks such as these might be used to encourage students to think about the prioritisation of causes and how to articulate this argument at the paragraph- and overall essay-level.

The key here though is that ‘the literacy’ is not a bolt-on. First, the language is introduced in the context of a wider causal enquiry and read in its original argumentative context so the students can see it in its natural habitat. (That is why the chunks below should not be considered an ‘off-the-peg’ resource because students studying other historiographies would be denied the opportunity to see how the historian actually deployed them.) Second, the teaching of language itself might make concrete to students that prioritisation is the driver of historical causal argument.[xiii]  Ultimately, the history teacher needs to continuously reflect on the possible ramifications of teaching literacy as an isolated ‘skill’.[xiv]  Otherwise, we might condemn students to procrustean structures that ignore the historical epistemology and contort, butcher, or kill students’ arguments.



[i] What follows is a critique of common sentence starters that I have either a) used or seen being used in my practice or b) seen shared on social media or sold on resource websites. I have not included images or links of the specific examples because I do not want to personalise any critique.

[ii] Pearson, Edexcel (2017). Advanced Level 3 GCE History, Paper 3: Theme in breadth with aspects in depth, Option 33: The Witchcraze in Britain, Europe and North America c.1580-c.1750 9H10/33 retrieved from

[iii]Carroll, J. E. (2017). ‘From divergent evolution to witting cross-fertilisation: the need for more awareness of potential inter-discursive communication regarding students’ extended historical writing’ The Curriculum Journal 28(4) pp.504-523

[iv] See, for example, Coffin, C. (2006). Historical Discourse: The Language of Time, Cause and Causation London: Continuum. p.126; Martin, J. R. (1992). English Text. System and Structure Philapdelphia, PA: Benjamins pp.454-456

[v] For example Counsell discussed the dangers of ‘bolting on’ literacy Counsell, C. (2004). Building the lesson around the text, history and literacy in year 7. London: Hodder Education p.4

[vi] Tosh, J. (2006). The Pursuit of History: Aims, methods and new directions in the study of modern history London: Pearson p.154

[vii] Evans R. J. (1997).  In Defence of History. London: Granta Books. P.158

[viii] Carr E. H. (1990). What is History? London: Penguin pp.89-90

[ix] Hewitson M. (2015). History and Causality. London: Palgrave Macmillan. p.161

[x] Hansen, C. (1988). Witchcraft at Salem. London: Arrow Books. pp.145-146

[xi] Hill, F (1996). A Delusion of Satan: The Full Story of the Salem Witch Trials. London: Penguin.pp.195-196

[xii]I describe this approach in far more depth here  Carroll, J.E. (2016) ‘The whole point of the thing: how nominalisation might develop students’ written causal arguments’ Teaching History 162 pp. 16-24

[xiii] This is essentially James Woodcock’s idea. See Woodcock, J. (2005). ‘Does the linguistic release the conceptual? Helping Year 10 to improve their causal reasoning’ Teaching History 119, 5-14.

[xiv]  See, for example, Evans, J. and Pate, G. (2007). ‘Does scaffolding make them fall? Reflecting on strategies for developing causal argument in Years 8 and 11’ Teaching History 128, 18-28.


Don’t you think we deserve an explanation? The limits of generic ‘explaining’ writing frames for constructing historical causal arguments



In a previous post, I argued against ‘sentence starters’ such as ‘I believe…’ being labelled as the ‘language of the historian’.[i] I now want to consider another manifestation of this phenomenon: ‘explaining’ writing frames such as;




First, let me specify what I mean by ‘historical explanation’. In everyday language ‘explain’ is often used synonymously with ‘elucidate’, ‘make clear’ or ‘explicate’. [ii] In historical discourse, however, ‘explain’ tends to have a more precise meaning. As Megill noted ‘for many historians ‘explanation’ particularly means to say what caused it.’ [iii] To ‘explain’ in history, therefore, usually means to provide an argument in response to a question focused on the second-order concept of causation.[iv] An example of such a question might be;

‘It was the unusual political conditions operating in Massachusetts in 1692 that explain the extraordinary events in Salem’. How far do you agree with this explanation of the Salem witch hunt of 1692? [v]

As this example shows, ‘explain’ here does not (only) mean ‘make clear’. Instead, the student is first expected to recognise the densely abstract cause given in the question (the unusual political conditions in Massachusetts in 1692).  Second, the student must identify further causes of the crisis that might be similarly abstract. Such causes could include: the trauma caused by setbacks in King William’s War against the Wabanakis and New French; the puritan elites’ anxiety at declining religiosity in their theocracy; or the legitimising role that powerful male societal gatekeepers played in endorsing the afflicted girls’ accusations. The student would then be required to argue which of the causes was most important in bringing about the witch-hunt.

One consequence of generic definitions of ‘explanation’ is that sometimes writing frames such as ‘…this shows that…’ are labelled as historical ‘explaining’ language. Such frames might be appropriate for ‘explication’ but I would argue they lack the specificity required to enable pupils to meet the particular demands of ‘historical explanation’ as defined above.

But even if we are more precise about focusing on ‘causation’ and winnow out ‘explicating’ language we still might encounter difficulties. Many historians in the historiography of Salem do occasionally use ‘so’, ‘therefore’, and ‘because’ for explaining causation. The trouble is that most of these historians use this language for different types of causation that are not specifically historical.

First, some historians such as Bernard Rosenthal might use such language to make clear their processes of logic. For example;

One need not press the issue that the kind of testimony offered against Burroughs was less than adequate for a civil libertarian; but one cannot easily let go of the discrepancy between the rigorous standards of evidence set by Increase Mather and his willingness to embrace the conviction of Burroughs. So we have something of a puzzle; why did Mather make such an eloquent case against spectral evidence, a case for strict criteria in the discovering of witchcraft, and yet conclude, on the basis of evidence that defied his criteria, that George Burroughs had been tried fairly?[vi]

Here, Rosenthal wanted to show the processes he went through in deciding the questions he wanted to investigate. Generally speaking, the influential puritan minister Increase Mather argued against the admission of spectral evidence (the notion than an afflicted person seeing an apparition of a ‘witch’ could be used as evidence that the accused had made a diabolical pact). Puzzlingly, however, Mather wrote approvingly of the conviction of the unorthodox minister George Burroughs despite the fact that it was mainly spectral evidence used to convict him. This conundrum therefore caused Rosenthal to investigate and identify what he considered to be a key cause of the Salem witch trials: a desire by the puritan elite to purge religious nonconformity. For our purposes, however, the point is that Rosenthal is not using ‘so’ here to argue the relationship between abstract historical causes and consequences but instead to lay out his chain of logic.

Second, when substantiating an abstract historical cause with evidence, some historians of Salem might use ‘so’, ‘therefore’ or ‘because’ in a fashion similar to ‘everyday’ usage: cause-and-effect between individual participants. For example;

When the girls spoke, adults listened, in part because they believed in witches and the girls seemed genuinely troubled, in part because Betty (Parris) was the minister (Samuel Parris’) daughter and Ann (Putnam Jr’s) kin nearly monopolised local office’.[vii]

Here, Peter Hoffer used ‘because’ to provide personalised examples of a broader, abstract historical cause: the unusual amount of credibility that societal gatekeepers gave to the girls’ accusations.  Before they start their secondary schooling, most pupils will probably already be practised in using ‘so’, ‘therefore’ or ‘because’ in this way for explaining causation in their everyday life – such as when they explain they don’t have their homework because their dog ate it.

While I do see a rationale for labelling ‘so’, ‘therefore’ or ‘because’ for the particular purposes described above as the ‘language of the historian’, I personally wouldn’t. This reticence is because I don’t see how using this explaining language to outline processes of logic or personalised causation is particular to academic history compared to other disciplines or everyday speech. That particular debate, however, is somewhat moot because history teachers often don’t model ‘so’, ‘therefore’ and ‘because’ for these goals.

I do also sympathise with the argument that in pupils’ early schooling the abstract nature of historical causation initially needs to be couched in the language of pupils’ lived experience. But a witting and grudging short-term compromise such as this is very different from claiming that using ‘so’, ‘therefore’ or ‘because’ for explaining historical causation represents ‘the language of the historian’. In fact, pupils will quickly need to discard such frames to explain historically. Ultimately, a deadening, homogenising and distorting effect will be the consequence of the longer-term use of frames such as;

There was a witch hunt in Salem in 1692 because…

Far from being the ‘language of the historian’, most historians from the historiography of Salem avoid using ‘so’, ‘therefore’, or ‘because’ in this way because it strangles argumentation at birth. Counter-intuitively, rather than using ‘connectives’ most of these historians instead often realise historical causal relationships in ‘uncommonsense’ ways by using prepositional phrases, verbs, and nouns.[viii] By doing so, these historians are able to argue regarding the role and relative importance of causes that essentially neutral words like ‘because’ do not allow. [ix] Consider the following examples;

With the loss of Maine and the ill-fated Quebec expedition, New England suffered its worst humiliation. In the tinder box of inevitable and angry recrimination, the witch hunt caught fire. [x]

The foundation of the witchcraft crisis lay in puritan New Englanders’ singular worldview, one they had inherited from the first settlers of Massachusetts Bay more than sixty years earlier.[xi]

There had been New England witch trials before, but none precipitated by a cohort of bewitched adolescent and preadolescent girls.[xii]

In these examples, Mary Beth Norton, Enders A. Robinson, and Stacy Schiff provided and characterised mostly abstract causes of the crisis: the desperation to identify scapegoats for the succession of humiliating defeats to the allied Wabanakis and New French (precondition); the suspicion of nonconformity in the puritan worldview (root); and the ‘afflicted’ girls’ accusations (precipitant). None of this characterisation is possible with ‘so’, ‘therefore’ or ‘because’.

As I hope these examples have demonstrated referring to ‘explaining frames’ as ‘language of the historian’ is often an inappropriate label. But what explains such frames being used in this way? As I argued here,[xiii] one reason might be that the frames and their recommended use are vestiges of centralised initiatives such as the ‘National Literacy Strategy in History’ (2002). [xiv] While this resource claimed to be history specific in reality much of the language provided was generically applicable to any ‘non-fiction’ genre. [xv] As I have tried to show, ‘so’, ‘therefore’ and ‘because’ are wholly appropriate in general explanations of processes of logic or in everyday speech. They might also have a place in disciplines in which causal explanations only involve isolated variables and therefore do not require characterisation or prioritisation (for example, perhaps, ‘the apple fell to the ground because of gravity’).

But ‘historical explanation’ has its own distinctive characteristics. Its ‘uncommonsense’ nature means that it will be very difficult for pupils to pick up its linguistic features without specific guidance. If we accept that schooling is, in part, designed to move students from the realm of lived experience toward disciplined modes of thinking then there seems little point in explicitly modelling the language of the ‘everyday’ longer-term. But the issue is deeper than that. Using language that is not fit for argumentative purposes has a stunting effect on one’s scope for historical argument. Ultimately, unless we model language genuinely rooted in historical discourse then pupils’ historical explanations will almost certainly have a deadening, artificial, and non-argumentative quality. Language intended well-meaningly by teachers as an introductory leg-up to the pupil might all too quickly become a shackle tethering them to the floor. On the other hand discipline-sensitive linguistic scaffolds might, to borrow James Woodcock’s phrase, actually serve to ‘release the conceptual’ and emphasise the importance of characterising and prioritising abstract historical causes.[xvi]




[i] What follows is a critique of common sentence starters that I have either a) used or seen being used in my practice or b) seen shared on social media or sold on resource websites. I have not included images or links of the specific examples because I do not want to personalise any critique.

[ii] Megill, A. (2007). Historical Knowledge, Historical Error: A Contemporary Guide to Practice. The University of Chicago Press: London p.79

[iii] Ibid.

[iv] Marwick, A. (2001). The New Nature of History: Knowledge, Evidence, Language. London: Palgrave.

[v] Pearson Edexcel (2016a). Specification Pearson Edexcel Level 3 Advanced GCE in  History (9H10) First Teaching from September 2015 First Certification from 2017 Issue 2 London: Pearson. p.133 Retrieved from

[vi] Rosenthal, B. (1993). Salem Story: Reading the Witch Trials of 1692. Cambridge: Cambridge University Press. p.142

[vii] Hoffer, P. C. (1996). The Devil’s Disciples: Makers of the Salem Witch Trials. London: John Hopkins University Press.  p.103

[viii] Halliday identified this phenomenon in academic writing which he termed ‘grammatical metaphor’.  Halliday, M.A.K. (1998). Things and relations. Regrammaticising experience as technical knowledge. In Martin, J.R. & Veel (Eds.) Reading Science: Critical and Functional Perspectives of Discourses of Science pp.185-237) London: Routledge; Halliday, M.A.K. & Matthiessen C.M.I.M. (1999). Construing Experience through Meaning. A Language Based Approach to Cognition. London: Cassell.

[ix] Martin, J. R. (2007). Construing knowledge: a functional linguistic perspective. In Christie, F. & Martin, J. R. (Eds.), Language, Knowledge and Pedagogy (pp.34-64). London: Continuum.

[x] Robinson, E. A. (1991). The Devil Discovered: Salem Witchcraft 1692. New York: Hippocrene Books p.189

[xi] Norton, M. B. (2002). In the Devil’s Snare: The Salem Witchcraft Crisis of 1692. New York: Alfred A. Knopf. p.295

[xii] Schiff, S. (2015).The Witches: Salem, 1692. London: Little, Brown & Company.

[xiii] Carroll, J. E. (2017). ‘From divergent evolution to witting cross-fertilisation: the need for more awareness of potential inter-discursive communication regarding students’ extended historical writing’ The Curriculum Journal 28(4) pp.504-523

[xiv] Department for Education and Skills. (2002). Key stage 3 national strategy – literacy in history. London: DfEaS p.21 Retrieved from

[xv] Ibid p.19

[xvi]Woodcock, J. (2005). ‘Does the linguistic release the conceptual? Helping Year 10 to improve their causal reasoning’ Teaching History 119, 5-14. Retrieved from


‘I believe…’ ahistorical ‘sentence starters’ should be scrapped. 


My research at the moment involves trying to identify and characterise the language academic historians use so that it might be explicitly taught to pupils. Consequently, I am particularly interested in the literacy displays or language mats I periodically see on social media or resource-marketplace websites that claim to be providing the ‘language of the historian’ or ‘how historians write’. I occasionally feel that such resources are at best inappropriately labelled because they sometimes include language that is directly antithetical to how most academic historians argue in writing. In this post I want to concentrate on the following ‘sentence starters’;[i]

I believe that…

I think that…

In my opinion…

Let me begin by outlining what I consider to be two essential and interrelated characteristics of historical writing based on theories of history.

1) As Fischer noted historical writing ‘always is articulated in the form of a reasoned argument’.[ii]

2) As Megill stated by definition a historical argument only exists when substantiated by evidence.[iii]

Most historians rarely use the ‘sentence starters’ above when referring to their own arguments because they directly contradict these essential characteristics of the historical epistemology. First, as Jordanova pointed out, ‘‘opinion’ is used when we want to draw attention to what a particular person thinks, without any particularly strong evidentiary base to back it up. Hence to call any bit of historical knowledge an opinion implies that it is not well-grounded, a merely personal view’.[iv] Second, ‘belief’ also has similar connotations. Consider the verb’s collocations in everyday language: people believe in Father Christmas, aliens or ghosts. The verb ‘believe’ is required here because there is no evidence to verify their existence.

Third, ‘belief’ is doubly problematical because in historical discourse it is often used synonymously with ‘ideology’. Most historians are often wary of the intrusion of belief/ideology into their peers’ arguments because it might lead to a scholar deliberately ignoring or distorting evidence in the furtherance of a particular presentist agenda.[v] This tendency also betrays the historical epistemology because, as Fulbrook noted, ‘historical knowledge is of a different order from that of fiction, myth and ideology’.[vi]

So, the  majority of academic historians rarely introduce their arguments with these sentence starters because it might insinuate that, historically speaking, what they are arguing is fundamentally defective. In fact, when historians in the historiography Nazism do use these phrases it is often when they are negatively evaluating a peer’s work. This evaluation is usually based on the criterion that the historian being reviewed has failed to sufficiently substantiate their claims. For example;

Mason thought he could detect clear signs, among these ordinary workers of ‘opposition’; if not political ‘resistance’, to the (Nazi) regime…Recent research has, however, begun to offer a different picture of the effects of economics upon working-class consciousness and behaviour during the 1930s.[vii]

This extract indicates that, in David Crew’s view, Timothy Mason made an overbold claim from the available evidence regarding the extent of workers’ immunity to Nazi indoctrination which has since been debunked by more recent research. Because the known evidence, according to Crew, now contradicts Mason’s assertion it can only qualify for the lesser classification of ‘thought’ instead of ‘argument’. Similarly;

By supporting this opinion with a questionable ‘primary source’ – the second-hand word of an anonymous Nazi doctor – Bock portrays women as complete victims and German men as either pseudo-victims or the primary perpetrators of Nazi terror.[viii]

Here, David Guba dismissed Gisela Bock’s claim as an ‘opinion’ due to her perceived questionable use of primary sources which, in Guba’s view,  undermined her attempted argument.

Also consider Gavriel Rosenfeld’s negative evaluation of William Shirer’s The Rise and Fall of the Third Reich;[ix]

From the outset of his book, Shirer made abundantly clear his belief that ‘Nazism and the Third Reich . . . were but a logical continuation of German history’…This was, to be sure, hardly a novel argument. Indeed, it marked a reversion to the notorious ‘Luther to Hitler’ view of German history, an overdrawn, deterministic perspective that was commonly espoused during the 1930s and 1940s.[x]

Rosenfeld criticised Shirer’s interpretative framework because, according to Rosenfeld, it led to Shirer reaching ‘overdrawn’ conclusions relative to the evidence. Furthermore, Rosenfeld went on to summarise how one of the main criticisms of Shirer’s work was that it was a thinly-veiled rehash of the tropes of Allied propaganda during and immediately after World War Two that claimed that Germans had become incorrigibly anti-democratic and militaristic at some point in their distant past. This ‘Vansittartist’ propaganda was partly designed to justify Allied reprisals, denazification and democratising initiatives in Germany immediately after the collapse of the Third Reich.[xi] This association of Shirer’s argument with propaganda helps explain why The Rise and Fall of the Third Reich was ‘bitterly attacked’ by West German critics for ‘being barely of elementary school sophistication’ and ‘distorting history’.[xii] Rosenfeld’s choice of the noun ‘belief’ might indicate here his subscription to the widely-held view that Shirer’s work was overly-driven by a political a priori thesis at the expense of misrepresenting the known evidence.

As I hope these examples have shown, a strange type of cognitive dissonance has developed in some secondary school history departments. Some pupils have been told that using ‘sentence starters’ such as ‘I believe’ to introduce their own arguments ‘is how historians write’, yet reading most works of academic history reveals that this is hardly ever the case.

I am not necessarily saying that all writing frames are bad. Nor am I suggesting that we should expect Year 7s to immediately write like Crew, Guba or Rosenfeld. I do want to offer two thoughts however. First, I want to preempt the potential ‘but-pupils-can’t-run-before-they-can-walk’ counter-argument. In my view the use of frames such as ‘I believe’ reflects an essentially faulty notion of progress. I have yet to see a convincing argument as to why drilling pupils with ‘sentence starters’ that are unhistorical in their early schooling is a necessary or even desirable step in pupils progressing to the construction of increasingly sophisticated historical written arguments later on.

The second thought might appear obvious. Words matter. Historians don’t select the words they use from the ether. Instead, the historian’s language often manifests tacit and commonly-held assumptions about what history is and how history works. It is for this simple reason that careful thought needs to be given by the teacher to the linguistic scaffolds they choose to give to their pupils. To take the example discussed above, in the ‘fake news’ era we might consider the repercussions of fundamentally misrepresenting the discipline to younger pupils. The conflation of the meanings of ‘opinion’, ‘belief’ and ‘argument’ might be why it is difficult to disabuse many pupils before they finish their compulsory history education of the flawed notion that, ultimately, there are ‘no wrong answers’ in history. It is perhaps not surprising that pupils develop this impression that ‘opinion’ and evidence-based argument are of the same essential order if they are explicitly and misleadingly told that ‘this is how historians write’.



[i] What follows is a critique of common sentence starters that I have either a) used myself or seen being used in my practice or b) seen shared on social media or sold on resource websites. I have not included images or links of the specific examples because I do not want to personalise any critique.

[ii] Fischer, D. H. (1970). Historians’ Fallacies: Toward a Logic of Historical Thought. New York: Harper Collins. p. xv

[iii] Megill, A. (2007). Historical Knowledge, Historical Error: A Contemporary Guide to Practice. The University of Chicago Press: London pp.6-7

[iv] The best summary of this distinction between opinion and historical knowledge I have found so far is from Jordanova, L. (2000). History in Practice. London: Arnold. pp.111-112 . Here is the whole extract: ‘It is often easiest to be precise about a concept by considering its opposites. The key concept in this chapter has been ‘historical knowledge’, so it may be worth considering what is not taken to be historical knowledge. Three terms are relevant here: opinion, ideology, and myth. ‘Opinion’ is used when we want to draw attention to what a particular person thinks, without any particularly strong evidentiary base to back it up. Hence to call any bit of historical knowledge an opinion implies that it is not well-grounded, a merely personal view. To call it ‘ideology’ is to make a somewhat different point, namely that it was driven by some prior commitment, such as a strongly held belief best understood as political, although here too a claim is being made about evidence. In the case of ideology, the evidence may well be there, but the charge is that it is being distorted or ignored because of the historian’s deepest assumptions’.

[v] Ibid.

[vi] Fulbrook, M. (2007). Historical Theory. London: Routledge p.196

[vii] Crew, D. (1995). General Introduction. In Crew, D. (Ed.) Nazism and German Society 1933-45 (pp.1-37) London: Routledge pp.2-3

[viii] Guba, J. D. A. (2009). ‘Women in Nazi Germany: Victims, Perpetrators, and the Abandonment of a Paradigm’. CONCEPT 33.

[ix] William L. Shirer (1960). The Rise and Fall of the Third Reich New York: Simon and Schuster

[x] Rosenfeld, G. D. (1999). ‘The Controversy That Isn’t: The Debate over Daniel J. Goldhagen’s “Hitler’s Willing Executioners” in Comparative Perspective’. Contemporary European History 8(2) pp.251-252

[xi] Hoenicke Moore, M. (2003). American Interpretations of National Socialism. In Rogers, D. & Steinweis, A. E. (Eds.) The Impact of Nazism: New Perspectives on the Third Reich and Its Legacy (pp.1-18) Lincoln: Nebraska University Press. pp.12-15

[xii] Rosenfeld op. cit. pp.251-252