These are the "relics and traces" (Marwick, 2001, 156) left by past societies, generated from within the society being studied.
Primary sources come about due to people living and organisations functioning (Marwick, 2001, 164). In other words, all documents are a testimony to the future about the present when they are produced. These are the documents an historian must find and use: it is what the historian does. Marwick's phrase is: Forget Facts, Foreground Sources (152-194).
Whatever one's historiographical position, there is surely a good argument beyond empiricism, about seeking out primary sources and being accurate to the sources as best as available. Otherwise there is loose cultural commentary at best or an over reliance of deductive theory without data, which in a sense never gets anyone past the proposals stage; writing remains reliant on secondary material, where the historian has already done the work. Looking again at a primary source means a fresh pair of eyes might find something missed or another angle from the material. So often secondary material is unreliable by omission or bias and to use it is to repeat and even expand upon an error. Its error over the primary sources is the laziness of the next historian.
A book on an historical period written in the twentieth century becomes a secondary source for its historical period, but it is a primary source for books written at that time (say in the study methods of historians and what it tells about the pursuit of history!).
Secondary sources usefully introduce a topic and show what historians have done. They give out areas of historical controversy. Yet there is a need (at postgraduate level and above certainly) to tackle primary sources - material of the time under study.
Primary sources contain secondary source material, as in comments on something else written, but it retains primary material in its direct recording. It also facilitates finding primary material it has referred to and referenced, sometimes quoting from primary sources (157, 161), whilst it remains important to go beyond extracts to these sources directly.
There are some essential steps of use in what Marwick calls The Catechism (182-184) points of which include the following:
An historian's approach should be in this order:
Here are different kinds of primary sources:
Primary sources contain:
The first is the document telling something about the time period that was built in but not intended. These include cultural assumptions and matters that the contemporary person hardly needs to mention but still includes unconsciously. They also fail to see the wider significance and so gloss over. The second is where the person is aware of the message inserted but does so wrapped inside the official message, and maybe the work of only some producing the document. Here the wider significance may be understood but it simply cannot appear too overtly. These approaches often come from oppressive regimes or from say religions where orthodoxy is valued. The third is the overt point of the document and, of course, it may attempt to partially or wholly mislead. (cf 172-174)
History can inform the arts and, with some care, the arts can inform history.
(Marwick, 2001, 187)
So the important thing is to treat cultural objects with care. Because they give off uncertain and subjective metaphors, cultural artefacts need context as they lack the provision of language.
Marwick, A. (2001), The New Nature of History: Knowledge, Evidence, Language, Basingstoke: Palgrave.
Pluralist - Liberal and Thoughtful