# Surmised attributes

**URL:** <https://discourse.citationstyles.org/t/surmised-attributes/277>\
**Category:** CSL Development\
**Created:** [June 24, 2007, 5:33pm UTC](https://discourse.citationstyles.org/t/surmised-attributes/277 "2007-06-24T17:33:41Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![John\_P\_McCaskey](https://avatars.discourse-cdn.com/v4/letter/j/6bbea6/32.png) [@John\_P\_McCaskey](https://discourse.citationstyles.org/u/John_P_McCaskey)\
**Post date:** [June 24, 2007, 5:33pm UTC](https://discourse.citationstyles.org/t/surmised-attributes/277/1 "2007-06-24T17:33:41Z")

</div>

Does CSL offer a way to flag attributes such as authors, publication dates,  
and publication locations as surmised or inferred? I didn’t see anything in  
the schema.

In my field of early modern history, such details are often not supplied on  
the book itself and must be guessed or inferred.

When they can be determined, such attributes are typically placed within  
square brackets, sometimes with question marks when the attribution is  
questionable

Here is a citation (or two actually) in a footnote from a book I’m currently  
reading (italics in uppercase):

[F. K. Hunt], “A Few Facts about Matrimony,” in Charles Dickens, ed.,  
HOUSEHOLD WORDS, 1 (1850), p. 374 (for authorship, see Anne Lohrli, comp.,  
HOUSEHOLD WORDS [Toronto, 1973], p.319).

A writer couldn’t just stick brackets in an author field, for the first  
book’s author should be “[Hunt, F. K.]” (and alphabetized without regard for  
the brackets) in the bibliography.

For the second book, if the date and location each had brackets, we’d get  
[Toronto], [1973]. (Even if a citation processor wasn’t smart enough to  
combine these, it seems the CSL should make such combining possible.)

John

---

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [June 24, 2007, 5:50pm UTC](https://discourse.citationstyles.org/t/surmised-attributes/277/2 "2007-06-24T17:50:06Z")

</div>

John,

> Does CSL offer a way to flag attributes such as authors, publication  
> dates, and publication locations as surmised or inferred? I didn’t see  
> anything in the schema.

No, and the thing is, this is an issue that ties together the styling  
language and data model. E.g. it’s not just a CSL issue.

In fact, if I was going to be lazy, I’d say it’s really not my problem;  
if someone wants to do an approximate date, it’s not CSL’s job to worry  
about how it’s represented.

I don’t believe that though, and would prefer to find a better way.

> In my field of early modern history, such details are often not  
> supplied on the book itself and must be guessed or inferred.

Yes, I know.

> When they can be determined, such attributes are typically placed  
> within square brackets, sometimes with question marks when the  
> attribution is questionable
> 
> Here is a citation (or two actually) in a footnote from a book I’m  
> currently reading (italics in uppercase):
> 
> [F. K. Hunt], “A Few Facts about Matrimony,” in Charles Dickens, ed.,  
> HOUSEHOLD WORDS, 1 (1850), p. 374 (for authorship, see Anne Lohrli,  
> comp., HOUSEHOLD WORDS [Toronto, 1973], p.319).
> 
> A writer couldn’t just stick brackets in an author field, for the  
> first book’s author should be “[Hunt, F. K.]” (and alphabetized  
> without regard for the brackets) in the bibliography.
> 
> For the second book, if the date and location each had brackets, we’d  
> get [Toronto], [1973]. (Even if a citation processor wasn’t smart  
> enough to combine these, it seems the CSL should make such combining  
> possible.)

As I say above, this is kind of a PITA. For example, on the data  
modeling side, we’d need:

```
- for Zotero, say, to be able to say when some data is inferred, 

```

opproximate, etc.  
- same for the RDF

But so how to do this? Is an inferred publisher in fact a different  
kind of relation? A different kind of datatype in the case of dates?  
E.g. in RDF N3 notation, are we doing:

[http://ex.net/1](http://ex.net/1)  
dc:date “-0534”^^[http://dates.org/inferredYear](http://dates.org/inferredYear) ;  
x:inferredPublisher [http://ex.net/2](http://ex.net/2) .

…? Or:

[http://ex.net/1](http://ex.net/1)  
x:inferredDate “-0534”^^xsd:gYear ;  
x:inferredPublisher [http://ex.net/2](http://ex.net/2) .

And how to deal with it in CSL? You’re right; CSL would at least have  
to understand the concept of inferred or approximate data.

I don’t think the answers are that straight-forward.

Bruce