The source of Sources
Sitting in the BNW – the brave new world, in Miranda’s sense, not Aldous’s – I read a phrase that caused an unbidden epiphany. Reading guidance to history undergraduates, I came across the term ‘secondary sources’ and I was involuntarily transported back in my mind to a late morning in the autumn of 1991, when I was bushy-haired and narrow-faced. I was sitting in St Hugh’s, the weak sunlight from Oxford’s Woodstock Road falling across the desk at which both my newly appointed supervisor and I sat. I remember George Garnett picking me up when I used exactly that phrase to refer to some historians I had been reading: ‘how can something secondary count as a source?’ I think were his words, looking at me through his thick-rimmed glasses with his characteristic look that could have been either quizzical or mocking (I was never quite sure which and that uncertainty lasts to today).
I blustered because I could not deny the logic: if a source is – as historians believe – the place from which information originates, then the work of an historian analysing that information cannot itself be the fundamental source. Indeed, if that latter-day historian is the source – if there is no prior, contemporary statement to the ‘fact’ – then there is something suspect about the information. So, since that supervision, I have avoided the solecism and, in my usage, taken care to distinguish between ‘primary sources’ and ‘secondary material’.
A quick look on-line, though, shows that George and I are in a minority. ‘Secondary source’ is the accepted phrase, especially in North American English with several universities, including Princeton, providing their students with a definition: ‘a secondary source interprets and analyzes primary sources’. We might justify this use by saying that a textbook or monograph may be someone’s source of information, even if it is not the origin of the information. We might also note that the terminology is certainly not confined to the States and Canada; it is there in John Tosh’s and Arthur Marwick’s textbooks on historiography, and it appears also in the Oxford Advanced Learner’s Dictionary. Life, though, is not as simple in the OED itself, where the phrase itself has no place but where there is a draft addition, dated September 2007, which talks of ‘secondary’ in a broad academic sense of ‘designating a text with another text as its subject’. That draft dictionary entry might alert us to the fact that there is a history to the development of the historian’s usage.
There is a history and also a peculiarity to the historian’s usage – the lawyer, for example, would talk of ‘secondary evidence’ as that which is supplementary to the primary or main evidence. My impression also is that, in historical writing in the nineteenth century, ‘secondary’ often had a similar meaning of supplementary or, rather, subsidiary. In another instance, I see a Victorian classicist referring to an ancient Syriac translation not being made from the Greek but ‘from some secondary source’, where the meaning must be ‘at one remove’ or ‘via an intermediary’. When such meanings become drowned out by present usage I have not ascertained – a brief glance of Collingwood’s The Idea of History (1946), for instance, suggests no engagement with the dichotomy we use, but that may say something about his own prejudice rather than common practice at that point – and would welcome your insight, gentle reader. My own train of thoughts have moved in a different direction.
Our speech is moulded by metaphors so deeply embedded that we fail to recognise our prose is often poetry. A ‘source’ is the origin of information is an eighteenth-century transformation of an older term with a clear primary meaning: the fountainhead of a spring, the point where water gushes from the earth. I am reminded of a short poem by Philip Larkin where he says that if he was called upon to construct a religion, he would make use of water. We historians are unwitting believers in that faith. But, if we are going to talk of where we find our evidence with aquatic metaphors, why should we use this particular sylvan imagery? We could talk of drawing our information from a well of knowledge, which might truthfully suggest that we take a mere bucketful when we could drink more deeply. Or might we refer to standing ponds or lakes and reservoirs to signify the mass of information – but we would have to recognise that to adopt such terminology would imply a concept of evidence as undifferentiated, gathered together into smooth conformity when we know quite the opposite is true, its conflicts, inconsistencies and downright incoherence being more like eddies, whirlpools and life-threatening weirs.
I press the point so we think more deeply about the dominant metaphor we ourselves use. Leave aside the oxymoron of a ‘secondary source’, a ‘source’ itself is problematic. No historian works with one vein of evidence and, when faced with multiple accounts of one event, is likely to be so struck by the incongruities that any similarities between them can seem suspicious. In its basic definition, a ‘source’ would provide a spring of inspiration that would widen into a stream and then join with other tributaries into a river – that sense of compatibility which is a symbol of consensus is surely not how we experience our encounters with the evidence. At the same time, though, there is perhaps one wonderful insight provided by our master, the metaphor: just as the evidence we have is not the ‘fact’ itself, but a witness to something (to what, though?), the spring’s source is a reminder that its own origins – the truth from which it bubbles out of the ground – lie subterranean.