historical source criticism

Works

A manual, a meditation, and the philological ancestry behind both.

• Langlois and Seignobos's Introduction to the Study of History (1898) is the codification — the external/internal division, and the insistence that no fact enters history without passing it.
• Bloch's The Historian's Craft (written 1941–43, published 1949) reframes the document as an unintentional trace: witnesses in spite of themselves are more reliable than witnesses trying to persuade.
• The philological tradition, from Valla's exposure of the Donation of Constantine (1440) onward, supplies the technique's origin — authenticity settled by internal evidence of language and anachronism.
• Modern archival theory extends the same questions to born-digital records, where provenance is a metadata problem rather than a paper one.

Central ideas

Four ideas, and the second reorders everything.

• External before internal: authenticity, date, authorship, and transmission are settled before content is weighed, because a misdated source produces confident nonsense.
• The document as trace: a record is primarily evidence that someone produced it, in that situation, for those reasons — and only secondarily evidence about what it describes.
• Witnesses in spite of themselves: incidental details, which the maker had no interest in shaping, are usually the most reliable content in a partisan document.
• Interest without deception: a source's stake shapes it below the level of intent, so establishing sincerity does not establish reliability.

Distinctive vocabulary

Five terms, precise in the discipline.

• Provenance: the documented chain of custody and origin. Not the source's claim about itself.
• External criticism: authenticity, date, author, transmission. Not judgment of content.
• Internal criticism: the author's knowledge, competence, and interest. Not fact-checking.
• Trace: an unintentional record. Bloch's category, and the discipline's preferred evidence.
• Collation: comparison of witnesses to a text to establish its transmission. Not corroboration.

Debates and disagreements

The discipline argues about what its own rigour excludes.

• Whether the method encodes a bias toward the documented: source criticism is exquisitely good at handling what was written down and silent about lives that left no records, which is the standing objection from social and subaltern history.
• Whether the external/internal division holds: dating often depends on content and content on dating, so the sequence is a working order rather than a logical one.
• Whether Bloch's traces are as innocent as they look: incidental details are shaped by what a genre made incidental, so the safe category is smaller than the doctrine suggests.
• Whether digital records fit: provenance for born-digital material is metadata that can be edited, and the discipline's paper-era assumptions transfer imperfectly.

Intellectual relationships

It is the practical wing of several things this record documents.

• Fallibilism, documented here, is the epistemology this construct implements: provenance and transmission are how a claim is made checkable by someone else.
• The historical-temporal lens supplies the theory of the archive that this construct's procedures presuppose — what can be known as past depends on what was kept.
• The media-theoretic lens supplies the hard floor: no record, no source, and no criticism of it.
• Standpoint epistemology supplies the sharpest objection — the archive preserves the perspectives that had the standing to be archived.

How it changes this reading

Seven placements, and this record is itself an archive being read.

• For profiles as question maps and machine interpretation as answerable, it supplies the procedural requirement: a claim carries its provenance, or it is a source with no criticism attached.
• For calibration language over characterological description, it supplies Bloch's rule — a document establishes that someone wrote this, then, under those conditions, and a claim about character is several unsupported steps beyond it.
• For architecture over perception management, it supplies one of the oldest instances of the preference: historians are not trusted to be careful, the apparatus is built so carelessness shows.
• For iteration, it supplies the archive's ordinary condition — a new find reopens a settled reading, and that is normal rather than failure.
• The self-implicating note: this corpus is a set of sources about one life, and everything the construct says about interest, purpose, and transmission applies to it first.

Useful comparisons

Against neighbouring methods.

• Against fact-checking: source criticism asks what kind of evidence a claim rests on, where fact-checking asks whether the claim is true — the first is prior and usually more informative.
• Against textual criticism: establishing what a text said against establishing what it is evidence for; philology is the ancestor and the narrower task.
• Against journalistic sourcing: overlapping instincts, formalized to different degrees, with the historian's advantage of not being on deadline.

Where the ideas meet

What the tradition holds throughout.

• Provenance precedes content: what a document is governs what it can show.
• Interest is structural: every source has one, and having one is not a defect to be corrected but a property to be read.
• Procedure over trust: the apparatus exists because individual care is not checkable.

Where they part

Where practitioners part.

• On the strictness of the sequence: external-then-internal as doctrine or as heuristic.
• On what counts as a source: documents only, or material culture, oral testimony, and absence as well.
• On the archive's silences: a limit to be acknowledged, or a subject to be studied in its own right.

Limits

What the construct cannot do.

• It cannot recover what was never recorded: the method's rigour applies only inside the archive, and the archive is a selection.
• It cannot establish content from provenance: knowing what a document is does not settle whether its claims are true.
• It cannot read intent: interest is inferable from position, and purpose from genre, but neither is the author's mind.
• It transfers imperfectly to digital records, where the chain of custody is itself mutable.

Criticisms

Standing objections, at strength.

• That the method's precision is available only where records exist, so its rigour systematically privileges the literate, the official, and the powerful.
• That the external/internal sequence is circular in practice, since dating and content inform each other.
• That its confidence in incidental detail underrates how thoroughly genre shapes what a maker treated as incidental.
• That professionalized source criticism has at times been a gatekeeping device against oral and community histories.

Common misreadings

Four.

• Sincerity as reliability: establishing that a source believed itself settles almost nothing about accuracy.
• Provenance as endorsement: a well-documented chain establishes what a thing is, not that it is right.
• Skipping external criticism on familiar material: the step most often omitted is the one that catches the largest errors.
• Treating absence as evidence of absence: the archive's silences are about the archive first.

What remains outside this idea

The boundary.

• That a document's claims are true: the method establishes what kind of evidence it is.
• That an author was honest: interest operates below intent, and sincerity is neither necessary nor sufficient.
• That what is not in the archive did not happen — the inference the discipline most consistently warns against.