A place-name survey example

Place-name surveys, cartularies and editions record things a simple list of places does not: when a manuscript was copied, where on the page a form is found, whether a form is a reading or an editor’s choice. This example shows how the spreadsheets hold them. It is loosely based on the English Place-Name Society’s surveys; the sources and readings are illustrative, so check an edition before citing any of them.

Download the example as CSV files.

The sources, and a manuscript copied later

source_id	title	citation	date	from	to	derived_from
asc-annal-921	Anglo-Saxon Chronicle, annal for 921	ASC s.a. 921	undated
asc-ms-a	A	ASC (A): Cambridge, Corpus Christi College, MS 173	c. 925	0915	0935	asc-annal-921
domesday	DB	Domesday Book, Buckinghamshire	1086	1086	1086
deed-c1200	Deed, c. 1200 (illustrative)		c. 1200	1190	1210
fine-c1240	Fine, c. 1240 (illustrative)		c. 1240	1235	1245
close-rolls	Cl	Calendar of the Close Rolls	undated
pn-bk	PN Bk	A. Mawer and F. M. Stenton, The Place-Names of Buckinghamshire (EPNS 2, 1925)	1925	1925	1925
deep	DEEP	DEEP: Digital Exposure of English Place-names	2013	2013	2013

Look at the first two rows. The survey’s notation “921 (c. 925) ASC (A)” means an annal for the year 921, read in the A manuscript of the Anglo-Saxon Chronicle, which was written about 925. That is two sources: the annal, and the manuscript, which is derived_from it and has its own date, c. 925, with from 0915 and to 0935. A manuscript’s sigil (“A”) is simply its title.

The names

place_id	name	language	script	form_status	occurrence_context	occurrence_count	date	from	to	source_id	locator	attribution	certainty	notes
cambridge	Grantanbrycg	ang	Latn				921 (c. 925)	0921	0921	asc-ms-a			1	The annal is for 921; the spelling is read in the A manuscript, written c. 925, whose date is on its source row.
bunsty	Bonestou					3	1086	1086	1086	domesday			1
bunsty	Bonestou				InPersonalName		c. 1200	1190	1210	deed-c1200			1	Willelmus de Bonestou, witness. The hundred is not named; a man from it is.
bunsty	Bonestou						1254	1254	1254	close-rolls	1253-54, p. 31	Inferred	1	Source given as ib. in the survey and resolved editorially.
bunsty	Bunstow(e)						c. 1240	1235	1245	fine-c1240
bunsty	Bunsty Hundred	en		Headword			1925	1925	1925	pn-bk				The editorial headword under which the survey files the entry.
bunsty	Bunstowe			Normalised			2013	2013	2013	deep				Search form derived from Bunstow(e); attested by no source.

Row by row:

• Grantanbrycg, dated 921 (the year of the annal) and sourced to the manuscript it was read in. The manuscript’s date lives on its source row, so the two dates never get confused.

• Bonestou in Domesday Book, found three times in the passage cited: occurrence_count 3, the surveys’ “(3 X)”.

• Bonestou again, but only inside a man’s name, Willelmus de Bonestou: occurrence_context is InPersonalName, the surveys’ “(p)”. This is certain as evidence, so certainty stays 1; what differs is the kind of evidence, and that is what the column records.

• Bonestou in 1254, where the survey gave the source as “ib.” and an editor worked out that the Close Rolls were meant: attribution is Inferred. The locator gives the volume and page.

• Bunstow(e), an ordinary reading, with the edition’s brackets kept.

• Bunsty Hundred, the form the survey files the entry under: form_status is Headword. It is an editor’s decision, not a reading, so it is dated to the survey and sourced to it.

• Bunstowe, a spelling made up for searching, attested by no source: form_status is Normalised, sourced to the project that made it.

Someone computing “the earliest spelling” or “how many independent spellings” can now leave out the headword, the search form and the personal name without having to know the survey’s conventions.

Types and relations

place_id	type_label	date	from	to	source_id
bunsty	hundred	1086	1086	1086	domesday

place_id	relation_type	related_place_id	date	from	source_id	notes
bunsty	ContainedIn	buckinghamshire	1086	1086	domesday	Domesday lists the hundred under Buckinghamshire; no end date is asserted.

The relation says that Domesday Book, in 1086, places the hundred within Buckinghamshire. It gives from but no to, because the source says nothing about when that stopped being true. Both places must be in the places sheet:

place_id	label	country_codes
cambridge	Cambridge	GB
bunsty	Bunsty Hundred	GB
buckinghamshire	Buckinghamshire	GB

What the spreadsheets could not hold here

The surveys also write runs such as “1252 Cl et passim to 1346 Harl”: one spelling attested continuously between two sources. That is a single piece of evidence resting on two sources, which a spreadsheet row cannot express; the JSON format can.