Hi Olivia,

Some resources:
Detailed elements of MODS
I’ve found the book “Metadata for Digital Collections” by Steven J. Miller invaluable.

If it causes a lot of confusion, the date qualifiers are optional… in my MODS data I actually have a text field date (non-sorting) where I model the date after RAD standards (Canadian form of DACS or ISAD). RAD standards are similar to DACS rule 2.4.15 for dates. I use the start dates and end dates to index dates. 

Do you use DACS? I would look at that rule and map that onto the rules for approximate, inferred, and questionable. 

Approximate to me means anything “circa” defined by the cataloguer, probable century/decade but uncertain exact year.
Inferred is a certain pinpointed date (before ——, after ——) defined by the resource
Questionable might be a date the resource or cataloguer defined, but for some reason is uncertain/unreliable

Hope that helps!

Tanis Franco
1265 Military Trail, Toronto, Ontario, M1C 1A4

On Dec 12, 2017, at 12:11 AM, Olivia S Solis <[log in to unmask]> wrote:

Hello all,

This is newbie question, but one I have been grappling with. FYI, I'm relatively new in my field. While I am learning and can generally understand cataloging standards, my strength as is data manipulation, not cataloging. Can anyone point me to some resources for defining "approximate" vs. "questionable" vs. "inferred" as date qualifiers? The definitions provided in MODS:

  • approximate – This value is used to identify a date that may not be exact, but is approximated, such as "ca. 1972".
  • inferred – This value is used to identify a date that has not been transcribed directly from a resource, such as "[not before 1852]".
  • questionable – This value is used to identify a questionable date for a resource, such as "1972?".

are helpful, but leave lingering questions. The distinction between the three isn't clear cut to me. I work at an archives and so am struggling with how to define these terms so that they are applied consistently at both the collection and item level. For instance (I'm going to use a mod example since this is the MODS list), say this image is in a collection in our archives in the Nancy Sinatra papers (which we don't have):

Since the includes Lee Hazelwood and Nancy Sinatra, I might infer (based on evidence in the photo) that it was a publicity shot taken around 1968 when their joint album was released. I might also infer that based on their clothing. If I didn't know anything about their joint album, based on their clothing and relative age (because I'm really familiar with Nancy as the collection's processor), I can also see this being approximately around the late 1960s (begin 1965, end 1969) since I cannot know the exact date it was taken. How questionable need a date be to be designated questionable?

I want some clear cut examples to give catalogers/processors, who might also have other evidence to draw from based on other items in a collection. Do any of you have more specific rules you enforce at your archives/library about situations to apply "inferred", "approximate", or "questionable"?

Thanks, and sorry if this question is very basic! At least at our archives though, I can tell this is an area of confusion. The three qualifiers have essentially been interchangeable in a lot of legacy data.


Olivia Solis, MSIS
Metadata Coordinator
Dolph Briscoe Center for American History
The University of Texas at Austin
2300 Red River St. Stop D1100
Austin TX, 78712-1426
(512) 232-8013