[In this second part of the chapter I develop the argument that history has the same features as language in general. It is incapable of being pinned down to a single factual original simply because we retell history in language. The way history is written - even at the most basic level - is mired in linguistic and other choices that have nothing to do with fact. Again, that is just how history is and the alternative is practically inconceivable anyway. 'Facts' and 'events' in history also have the same features as other units of language aimed at conveying information; they all have their signifieds, and are all capable of infinite repetition so that, again, there is no fixed originary meaning. In conclusion I repeat that none of this renders history detached from the importance of empirical accuracy, and is no licence to make things up or to deny that things happen. It is simply the way that history is, but we need to emrace the possibilities that that presents - which is what the second half of Why History Doesn't Matter is about.
The same caveats apply as for part one and, for the same reasons, the notes start at 'iv.'..]
Historical narrative is “structured like a language”
Words matter. It is of crucial importance, if we are to explore how History can create a role for itself separate from simple chronicling and antiquarianism, to develop the point about the literary or rather linguistic nature not just of historical writing but of historical narrative itself. Even the dullest chronicle is trapped in a linguistic net that makes it much more than a simple description. The vocabulary we use to describe past events never bears a simple or direct to the historical object being discussed. Let us take as an example a seemingly uncontroversial description of a well-known event: The Battle of Waterloo (18 June 1815). To name it the Battle of Waterloo is already a decision to accept the name given to the encounter by the victors, rather than the (now much less common) French name of La Belle Alliance. To call it “a British victory” (let alone an English victory!) is to make several more decisions. First, one has ascribed the victory to the British, rather than “the Allies”; although the army might have been commanded by the Duke of Wellington, his Dutch and German allies made up 60% of the army.[iv] Even to label the battle an allied victory would be controversial to those who see the intervention of the Prussians as the decisive element. That, since 1815, convention in anglophone history has passed from referring to the battle as an English victory, through ascribing the success to the British, to acknowledging that the battle was won by allies drawn from numerous nations – leaving aside those who wish to call it a Prussian, or even “the German”,[v] victory – is proof enough of the baggage that even seemingly straightforward descriptions can carry. More importantly, though, all those options still use the word “victory”, implicitly therefore categorising the event from the perspective of the winners, be they British, allied or Prussian. It is a meaningful difference from the choice to describe the engagement as a French defeat. None of these various permutations of vocabulary – and we have left aside more obviously contentious language, such as describing it as a triumph or a catastrophe[vi] – is entirely neutral. Apart from the description of the victory as “English”, each can be argued to work perfectly well as a description of the events of 18 June 1815. On the other hand, those events themselves impose no specific vocabulary on the historian. Even a phrase as banal as “on 18 June 1815 the Allied army won the Battle of Waterloo” is thus already tangled up in a network of linguistic choices that take it away from bearing an unmediated relationship to the object described. To escape the latter, one would have to describe the day in something like the following manner:
On the date identified in conventional calendars as 18 June 1815, several hours of heavy fighting took place around various farms and hamlets to the south of Brussels, at the end of which the army comprising troops from Britain, the southern and northern Netherlands, Prussia and various other German states moved forward towards the south and the army largely but not entirely composed of French soldiers fighting for Napoleon Bonaparte moved back away from them and towards Paris after losing more men killed and wounded.
No one can seriously envisage any kind of history of the Napoleonic Wars being written in that kind of laborious prose! To the purveyor and consumer of writing that chronicles history’s many interesting tales, this sort of issue is of no more than stylistic importance[vii] or interest at best, and rightly so. However, minor though it may seem, this point lies at the heart of what I will suggest are the potential reasons why historical research could matter. The choices that I described above are in some ways all the more important for being barely conscious ones relating to a comparatively banal piece of labelling. Even they have political or other connotations; the decisions are those of the writer; they are not empirically imposed by the evidence itself. As we move further into the nature of historical writing, the results of such decisions become more significant still.
Here we must consider some philosophy, which will be expanded, and its implications developed, later. This might look like a detour from the subject of this chapter but it will be vital for my argument and it needs to be introduced here for reasons which will become apparent. In elementary linguistic theory, the letters D-O-G are obviously enough not actually, in and of themselves, a dog; what linguists call the sign (e.g. the word “dog”) is a fusion of the signifier (in our example, the letters D-O-G) and the signified (here the concept of a barking, tail-wagging quadruped). That speakers of a language agree on that sign is something that has developed over centuries and which continues to evolve. That raises the absolutely crucial point that to carry meaning the sign must be “iterable” (that is to say repeatable in any situation involving speakers of a language and the idea of a dog). This applies to any system of communication, even if devised as a private language between two siblings and no one else. If I tell you that I saw a mnyargle going past the window, you won’t have the faintest idea what I am on about, because I just made the word up. If I explain that a mnyargle is a furry, tail-wagging, barking quadruped, however, you will understand. Now that you understand, though, I can go on talking about mnyargles to my heart’s content, albeit at the expense of you thinking that I have lost my mind. More importantly, though, you can pass this book to someone else, who can read this passage and thereby also know what a mnyargle is, or you can leave the book on the bus, it can be picked up by someone else and they too can understand the meaning of mnyargle, even though neither of us is present. Someone can find a copy of this book decades after my death and yours, discover what the word means and use it correctly. This issue does not only, as was once generally assumed, apply only to writing but to speech too (substitute “recording” for “book” in the example above and the point remains the same). Indeed, all signifying systems, whether spoken and written or not, share these same features. If I see a dog I understand what it is because of an iterable concept or category of what a dog is.
Because of all that, there is no point where a particular word or term relates to a specific object in a pure, unadulterated form. Even in a hypothetical situation where I live alone on an island and have never spoken to anyone else (how that is possible need not concern us; it’s like Tarzan), the first time I see a dog, and understand it as member of the category “dog” (rather than as “friendly animal” or “threatening animal” or “Mr Woofy”[viii]), logically that category must already be in my mind and does not exclusively refer to that specific beast in that particular encounter. Logically (if not temporally) the concept precedes its first use. This is because all words convey their meaning by virtue of their difference from all the other words or categories in the language. So, it is not simply the case that D-O-G = furry, barking, tail-wagging quadruped, but also that D-O-G ≠ Cat (or pig, or iguana, or hat, or…). The preceding discussion, and especially the points about the radical separability of signifying systems and their particular users, summarises, as some readers will have noticed, the ideas of the still unjustly-maligned and calumniated French philosopher Jacques Derrida (1930-2004). Invoking Derrida’s name can act as something of a shibboleth, which is why I have held off from mentioning it until after I have described the key points of his theory.
Returning to the latter, in sum the meaning of a word is perpetually deferred – you can never get back to its pure, originary meaning – and stuck in equally endless chains of difference.[ix] For this reason, Derrida coined the word différance[x] to combine these features and describe this essential feature of language. Again, the point applies to every sign in every signifying system. They all function, in Derrida’s terms, “textually”, in that they share the essential features of text, with meaning conveyed by spacing, juxtaposition, différance and so on. Thus, when Derrida said that there is nothing outside text this was what he meant; you can never get outside différance; you can never get access to some sort of “beyond-text” (hors-texte) where meaning is pure and absolute.[xi] This in some ways is acknowledged in the famous beginning of Saint John’s Gospel: “In the beginning was the word [Gk logos; Latin uerbum]. And the word was God and the word was with God.”[xii] The point where meaning is absolute and outside différance can only be the divine. Outside the sphere of the theological, though, there is always the potential for slippage. In this fact lie the linguistic resources for irony, sarcasm and pun. After all, the word ‘d-o-g’, with which I began can also be a verb, but let us leave that there. I would not want dogged discussion of dogs to dog our steps through the rest of this chapter.
More importantly still, there are always points where writers make jumps, in their choice of words or concepts, that are not governed, empirically, by the nature of the thing they are discussing. In the Waterloo example above, these would be in the decision to refer to the engagement as a victory or a defeat. Where we can identify those points, which Derrida called “aporias”, we can open up points of decision that allow us to look at the writer’s assumptions and at alternative meanings or readings. Again, this is not an issue that is important to the consumer of public chronicling or antiquarianism, and that is fine. I will gradually present an argument, however, about why it is essential to making history matter.
For now, though, I would stress that this throws the problems with the idea that there are “true stories” that students can be taught into yet starker relief. The sorts of problems I have just briefly outlined saturate any narration, not only in the choice of words but also in the choice of content. The points that Derrida made about “text” can be seen to apply equally well to events. After all, within a historical narrative it is not only the case that “the Battle of Marston Moor” = “the events that took place on 2 July 1644 between the villages of Tockwith and Long Marston, north of York”, but also that “the Battle of Marston Moor” ≠ “the Battle of Naseby” (or “the battle of Edgehill”, or “the Battle of Lostwithiel”, etc.). The “does not equal” sign in the previous sentence encompasses the relative chronology, the differing preceding campaigns, subsequent consequences and military contexts, different tactical outcomes and so on. All those differences help constitute the things that identify Marston Moor and thus, within historical narration, make up the “signified” of the “signifier” “Battle of Marston Moor”. Even the notion of “the battle of X” is iterable. After all, staying within the mid-seventeenth century, there were two “Battles of Breitenfeld”, two “Battles of Nördlingen” and, returning to England, two “Battles of Newbury”. The point is perhaps driven home by the fact that there were also two “Battles of Lützen”: one in 1632 during the Thirty Years’ War and one in 1813 during the Napoleonic Wars. Unlike the battles listed earlier, which all took place during the same war and thus at least get numbered by historians for convenience, these are both simply called “the” Battle of Lützen. If I were to tell you that one of my ancestors was killed at the battle of Lützen, you might go away impressed that I could trace my ancestry back to the 1630s, when I was only talking about 1813. Within a narrative, then, events take on a function similar to those of words within a sentence or passage. They are chosen and placed to convey meaning according to positioning and juxtaposition with other events. We have seen this in the example of the “Northumbrian feud” and the sacked diarist discussed earlier, as well as in the events chosen for depiction in the films about the Zulu War. Just as aporias are encountered in authors’ choice of words, they exist equally in the choices of the events that make up a narrative. Those structural or functional similarities also permit the same kinds of slippage or miscommunication, so that a story can be taken quite a different way from that intended, as was seen earlier.
As well as the implicit spaces in historical language between words (or signs) and meaning, and between what is said overtly and what could just as easily have been said, there are other gaps that can be opened. These are the ones closed in the processes of selection and narrative. As was shown in the example of the “Northumbrian Feud” events are linked by juxtaposition. Placing one episode directly after another, a causal link is suggested even without necessarily using overt vocabulary like “as a result”, “consequently” or “therefore”. The time between the events, and everything that happened during that time, are closed up or covered over in such juxtaposition. The Durham Anonymous closed up spaces of about ten, at least seven and thirty-six years by placing one after another the events that he wished to link in a chain of cause and effect. This juxtaposition and the closing up of time that it involves are vital in giving events their precise significance and signification. When Cy Endfield opened Zulu with shots of the devastation of the British army’s camp at iSandlwana, replete with dead redcoats, he was building a particular story. Endfield and John Prebble, who co-authored the screenplay, also – shamefully – made out that the Zulus launched their attack on the British while pretending to discuss peace (the diametrical opposite of the events of late 1878 and early 1879). So, having swept aside the main British army in their attack, the Zulus were storming on towards Rorke’s Drift, where only 100 British soldiers stood in their way. Keep watching to see what happened… Fifteen years later, in Zulu Dawn, Hickox worked from a screen-play by Endfield, who perhaps wanted to make up for the distortions of his earlier film. The prequel opens the temporal space closed by Zulu, to recount how the British troops at iSandlwana came to find themselves being slaughtered and their camp destroyed. Telling a story of British imperial and aristocratic arrogance, the immediate background to Rorke’s Drift, and thus the battle itself, take on quite another appearance. Watch Zulu directly after watching Zulu Dawn and you are likely to “read” it rather differently from the way intended by its writer and director. Again, it must be said that this does not only apply to screenplays or fiction. ‘[T]he narrative of the barbarian invasion and settlements can be said to have begun’ in 376, says Ian Wood.[xiii] Indeed it can, but for not any particularly decisive reason. Historians all have to choose starting- and end-points for their work, almost always out of convenience. Without wanting to embroil the reader in a technical discussion of late Roman History, the narrative of the barbarian migrations could as easily start in any number of dates other than 376. People (including me)[xiv] have tended to start in 376 for several reasons including simple historiographical custom. One reason for that traditional starting date, however, is the assumption that the Goths who sacked in Rome in 410 and went on to settle in the south-east of what is now France in 418/19 were the same people who had crossed into the Roman Empire in 376 and defeated the Eastern Roman army at the battle of Adrianople (378). It is assumed that they remained an effectively unsubdued “people” within the Empire, who eventually rebelled again under Alaric in the 390s and the rest, as they say, “is history”. But the traditional narrative, like the story of the Northumbrian Feud, closes up time, in this case a decade or so between the end (in 382) of the Gothic crisis that erupted in 376 and the outbreak of the Gothic rebellion under Alaric in the mid-390s. If that closed space is recognised and opened it is possible to argue that the resolution of that crisis, in fact marked the end of any meaningful story of the Goths who crossed the Danube in 376 as a distinct group of people.[xv] On the other hand, opening the space closed by starting the narrative in 376 can, in the same way that Zulu Dawn does for Zulu, find a prequel that casts the later story in quite a different light. In this case we could see the origins of the Gothic crisis not in the appearance out of the blue of the Huns, a historical deus ex machina, but in the destabilising of the region north of the Danube by Emperor Valens’ Gothic War of 367-9.[xvi] We could even take the story right back to a treaty between the Goths and Emperor Constantine I in 332.[xvii] In some ways one critical task of the historian is to create innumerable ‘prequels’ to other historical narratives.#
In this chapter, I have attempted to illustrate the relationship between time, experience, record and the creation of history and to suggest the literary and, especially, linguistic features that saturate the latter. Any meaning that events have depends upon their place in a retrospectively constructed narrative. It depends on their juxtaposition with the other events selected within the narrative and upon the type of narrative that is told. All that depends entirely upon the contingent attitude of the historian/narrator/rememberer. None of that affects the relative historicity or the facticity of the events recorded in the sources. Let’s be clear: what I have said is not a license to deny that things occurred. The past, as the unity of all time, thought and action up until this very moment, here, now, happened and cannot be changed. But, as I have shown, it is absolutely meaningless in and of itself. It only takes on meaning through the way its contents are selected and arranged in the present. That history is something that can be and is changed. Regularly. To uncover the true story of history is to discover that no single “true story” ever happened and that none of history’s myriad stories is, in itself, “true”. And any account that demands to be accepted as the true story or the national narrative is almost certainly myth.
[iv] See above, p.000 and p.000, n.000
[v] Peter Hofschröer, Waterloo: The German Victory
[vi] Note that at the start of this chapter I deliberately described iSandlwana as a “disastrous British defeat” rather than as a “triumphal Zulu victory”. The latter would have made no sense given the point of the story that I wanted to tell.
[vii] As for instance in the way I have used “battle”, “encounter” and “engagement” as descriptions of Waterloo to avoid repetition of the same word.
[viii] Jacques Derrida argued, correctly, that even personal names are ultimately subject to the movements of différance. Derrida, De la grammatologie
[ix] The best analogy is that dictionary definitions only really point you at other definitions and so on, ad infinitum.
[x] Pronounced exactly the same as différence. The fact that the difference between the two words is only detectable in the written form but nevertheless conveyable in speech is deliberate, to underscore Derrida’s point about the fundamental equivalences of writing and spoken language.
[xi] Derrida De la grammatologie p.227 (Of Grammatology p.158): Il n’y a pas de hors-texte (there is no “beyond- (outer-, or outside-) text”. Typically, rather than make any attempt to engage with this point, Derrida’s “analytic” detractors have generally just retreated into their usual lazy, ill-informed traduction and mockery. If there’s nothing outside text how can you be hit by a bus? S. Blackburn, Truth. A guide for the Perplexed (London, 2004), p.170. Ho ho. The response of course is that if the bus exists outside a network of meaning and différance, how would you know to get out of the way? Blackburn, another of the self-appointed policemen of analytical rigour, had never actually read Derrida, much less tried to understand his point. If he had, he would have known that “il n’y a pas de hors-texte” did (and does) not mean “there is nothing outside the text” as he claims.
[xii] John 1.1
[xiii] I.N. Wood, ‘The barbarian invasions and first settlements’, in The Cambridge Ancient History, vol.13: The Late Empire, A.D. 337-425 ed. Cameron, A.M., & Garnsey, P., (Cambridge, 1998), pp.516-37, at p.517.
[xiv] G. Halsall, Barbarian Migrations and the Roman West, 376-568 (Cambridge, 2007),
[xv] Halsall, Barbarian Migrations and the Roman West, 180-185.
[xvi] Halsall, Barbarian Migrations and the Roman West, 170-75
[xvii] As, e.g., P. Heather, Goths and Romans, 332-489 (Oxford, 1991).