Protocols for Name Linkage

The assignment of a Person_ID number in MLD is done only when there is a high degree of certainty that people with the same (or similar) name are indeed the same person, or that a name recorded in a source is really an alias for another person. For entries from MASL and A. P. Beaven’s Aldermen of the City of London, Person_ID numbers were assigned based on the name linkage data offered by Anne Lancashire and A. P. Beaven.  For the Mercer biographies, we followed the name linkage developed by Anne Suton. But such assignments are more difficult for less politically prominent people, so a series of protocols have been developed to guide editors and cataloguers in assigning Person_IDs. In other words, how do we determine which John Clerk a record for ‘John Clerk’ belongs to? In what follows, ‘London’ also refers to Souhwark and Westminster. [To cite this page: M. Kowaleski, “Protocols for Name Linkage,” Medieval Londoners Project, https://medievallondoners.ace.fordham.edu/protocols-for-name-linkage-1/, version 24 June 2020]

A. To make a name linkage, all of the following (1-5) must be true:

1. There must be some evidence that the person resides in medieval London: for example, deliberately stated as ‘of London’ or involved in activities (such as city watch, office-holding, jury service, etc…) typical of residents. Owning property in London is suggestive but by itself is not an indication of residence unless the person is said to inhabit or reside there. AND
2. The Forenames must be the same (be aware of the English versions of Latin names) AND
3. The Surname must be the same; but here cataloguers need to be aware that:
(a)  y and i are usually interchangeable and, on occasion, so is eu and w, and ou and u, and
(b) minimums (n, u, v and m, ni, in, ui, iu, vi, iv, etc..) can easily have been mistaken by a previous transcriber
(c) Latin case endings are often indicated for show or by an abbreviation that should not be transcribed in the English forename and surname
(d) that Middle English forenames and especially surnames can have very wide variations in spelling (so that making sure that the surnames sound the same is as acceptable a match as exact spelling)
(e) Sound-alike letters such as s and z (in Boson, Bozoun) can be interchangeable; f and v are also sometimes interchangeable
(f) the ‘de’ of locative, ‘atte’ of topographical, and ‘le’ of occupational surnames often disappear over time from the surname; thus John le Baker in 1296 could be the same person as John Baker in 1304 (as long as there are other factors linking this person)
(g) Latin documents sometimes use Latin word forms for surnames that later appear in English spellings, such as Pistor for Baker, Faber for Smith, etc…
4. If the Surname is not the same, there must be a direct reference linking an explicit alias to the Surname within 10 years of the two surnames appearing in the records AND the context must also promote name linkage (for example, both appear in a deed referencing the same property, or in the same gild or same occupational context (such as a list of those fined for forestalling), or of a similar political status, or residence in the same ward or parish, etc…) AND
5. There must be no contrary evidence that the records refer to the same person. Contrary evidence (also called deal-breakers) includes:
(a) The date of death (usually indicated by a will) of the person comes before the record’s date
(b) The craft or occupation of the person is different (transfer between gilds was possible, but we try to keep track of these occurrences)
(c) The names of the parents, spouse, and children do not match (here the frequency of second and even third marriages need to be kept in mind)
(d) The status (as indicated by titles, citizenship, civic office, gild office, tax paid, property ownership, and ethnicity) does not match (although here the possibilities of gaining status over time must also be considered)

B. People in records separated by more than 10 years will not be linked unless:

6. The date gap between the two records (or the date range of the records in question) is less than 50 years AND at least two of the following protocols (7-9) are true:
7. There are records (such as wills, property transfers, court cases, etc…) that fill in the gaps between the ten years that show that the person was likely alive in the period between the two records. Since few medieval people had adult lives lasting more than 50 years, records that together span over 50 years for (seemingly) one person are to be treated suspiciously unless the early records demonstrably refer to the person as a child or adolescent;
8. One of the following characteristics are shared between the two records: occupation, craft, craft_office, parish of residence, ward of residence;
9. The recorded names being linked can be shown to share relatives (spouse, children, parents, or other kin).

C. The biggest problem is distinguishing family members (usually men) with the same name. They could be father, son, and grandfather; uncle and nephew, and even brothers (it was not unknown for a family to use the same forename for two or three of their sons). Here it is important to be especially sensitive to possible life spans, date of death, and spouses’ names. Medieval use of ‘sr’ and ‘jr’ or ‘the elder’ and ‘the younger’ or adding an occupational (John Smith the baker vs John Smith the fuller) or locative identifier (John Smith of Aldgate vs John Smith of Cheapside) tend to be used in medieval records to help distinguish those with the same name. In MLD, we use I, II, III, etc… to distinguish individuals with the same name (the Roman numerals are assigned as they come up, so they do not reflect chronological sequence)