Wikidata talk:WikiProject Historical Place

Welcome to the discussion about historical place!

Use the button in the top of the page or next to the title of this comment to add a new topic. Sign your comments by adding four tildes (~~~~). Use indentation in the comments by adding an extra colon (:) in the beginning of each paragraph you write. If you don't have a user account, please register to get one!

A sequence of items vs. changing properties of one item

Susannaanas (talk) 07:40, 29 March 2016 (UTC) Abbe98 (talk) 16:53, 29 March 2016 (UTC) Iberti (talk) 17:05, 29 March 2016 (UTC) Eetu Mäkelä (talk) 18:13, 29 March 2016 (UTC) Jura. Good idea! Interested in place names (mainly). Humphrey Southall (talk) 13:43, 11 April 2016 (UTC) Fralambert (talk) 22:36, 11 April 2016 (UTC) Rainer Simon (talk) 10:10, 17 April 2016 (UTC) Melderick (talk) 13:56, 7 June 2016 (UTC) Llywelyn2000 (talk) 16:47, 7 June 2016 (UTC) Vladimir Alexiev (talk) 10:36, 15 June 2016 (UTC) Capankajsmilyo (talk) 21:04, 1 November 2016 (UTC) Ainali (talk) 08:59, 24 November 2016 (UTC) Jheald (talk) 01:15, 17 February 2017 (UTC) B20180 (talk) 05:33, 27 July 2017 (UTC) Elinese (talk) 19:19, 23 March 2018 (UTC) Salgo60 (talk) 10:34, 13 April 2018 (UTC) PKM (talk) 01:31, 15 April 2018 (UTC) Tris T7 ^{TT me} Tris T7 (talk) 19:53, 29 January 2019 (UTC) Sp!ros (talk) 16:07, 3 April 2019 (UTC) Pollockc (talk) 18:26, 25 November 2019 (UTC) ChristianSW (talk) 14:10, 10 December 2020 (UTC) Stephen Gadd (talk) 10:31, 12 January 2022 (UTC) Iain Hallam (talk) 01:36, 5 April 2022 (UTC)

Notified participants of WikiProject Historical Place: I invite you to discuss the key principles about modeling a historical place. I thought we could start by drafting guidelines for the choice between a single place item with changing properties and a sequence of different items over the course of time, see here. Please bring in any examples, extra information, or introduce concepts that we can use in the discussion, if needed. Also, please invite anyone who you think should take part to join! --Susannaanas (talk) 12:00, 12 April 2016 (UTC)

Historical regions

At Wikidata:WikiProject Historical Place#Historical regions, I added an approach I suggested to Llywelyn2000 for building lists of people/places located in historical regions.

SELECT ?item ?itemLabel ?yob ?yod 
WHERE 
{
	?item wdt:P19 ?pob .
	?pob wdt:P131* ?parts .
	wd:Q104285 wdt:P527 ?parts .  
	OPTIONAL { ?item wdt:P569 ?dob . BIND(YEAR(?dob) as ?yob) }
	OPTIONAL { ?item wdt:P570 ?dod . BIND(YEAR(?dod) as ?yod) }
	SERVICE wikibase:label {  bd:serviceParam wikibase:language "en" }  
}
LIMIT 2000

Try it!

A query like above could find people born there. Obviously, a date-range could be added.
--- Jura 07:55, 27 May 2016 (UTC)

UK

For the UK, there may be the chance of a big data import from the Vision of Britain to populate external identifiers Vision of Britain unit ID (P3615) and Vision of Britain place ID (P3616). VoB has a well-developed structure for the different historical units and hierarchies into which the country has been divided up, used to ground historical information and many historical statistical series. The information on these relationships may also be available.

Because it is particularly interested in these units as units to locate statistical data on, and aggregate data between, the site keeps the different hierarchies in strictly separated silos, summarised here:

• hierarchies

But that may not fully reflect our purposes.

The data seems to raise two of the classic questions highlighted on the main page:

• in some cases there are entities or territories with an apparent historical continuity persisting between different hierarchies. (The key classes in which such identifications may have the potential to be made are summarised on the page: successions). To what extent does it make sense to model these here strictly as separate items; or does it sometimes make sense to model them here as a single item ?
• to what extent should we allow located in the administrative territorial entity (P131) to "read across" between different hierarchies ?

...more in due course. Jheald (talk) 20:54, 17 February 2017 (UTC)

Picking this up again, there are a further couple of issues I'd like to add that also seem to me foundational:

how to query hierarchies that depend on qualifiers

(Something I've also just raised on the queries talk page; but we'll see whether anyone comes up with any suggestions).

It's familiar and straightforward enough to try to see whether something was in a particular subtree of a hierarchy, or to enumerate all the items in that part of the hierarchy, using queries with wdt:P361* and wdt:P131* for part of (P361) and located in the administrative territorial entity (P131) respectively.

But can one do a similar thing and take account qualifiers on those item-wise statements? For example the statements might be subject tostart time (P580) or end time (P582) qualifiers, or perhaps some qualifier representing "applies to hierarchy" H.

As entities get split etc, boundaries get changed, the membership of one of these sets might be somewhat different for one year, that (say) for twenty years later. The information should (could) be there in the qualifiers, but can we write queries that take account of them?

Do we have to give up on the easy recursiveness of wdt:P361*, and explicitly consider each link one at a time?

Are there ways to extact a pruned hierarchy, then run queries on that? (eg perhaps something like a Named Subquery, but perhaps extracting a graph rather than a table, does Virtual Graphs do something like this?) Jheald (talk) 13:59, 20 February 2017 (UTC)

Update - I don't think "Virtual Graph" helps. As far as I can see, this is just a convenience function to allow several graphs that are already defined as named graphs within the triple store to be given a collective name and used together.

As for named subquery, that could be useful to extract the allowed items once (as a set), and for various operations to then be run against that set.

But what might work would be to extract candidate statements as a set, then restrict to allowed statements, then find paths through that. No idea whether it would be efficient though. Need some test cases, too. Jheald (talk) 16:15, 20 February 2017 (UTC)

The idea of first running a sub-query to identify "good" P131s, then somehow reusing it in different contexts in the main query, maybe isn't viable: a query for P131s in Scotland for 1980 tinyurl.com/grxfh57 completes in 2 seconds and finds 11,000 such relations; but the equivalent query for England tinyurl.com/h7apvvu times out. Jheald (talk) 19:36, 20 February 2017 (UTC)

deviations from strict nesting

I am thinking of cases where A is in B; B is partly in C and partly in D (a deviation from strict nesting); but A is all in C and not at all in D

For example North East Lincolnshire (Q1342914) is part of (P361) Lincolnshire (Q23090), but located in the administrative territorial entity (P131) Yorkshire and the Humber (Q48063) -- however some parts of Lincolnshire are located in the administrative territorial entity (P131) East Midlands (Q47994)

So even before we get into different historical variations; or combining multiple hierarchies, and potentially trying to read across them, how do we deal with this in the even the simplest case, of non-nesting in a single hierarchy, at a single point in time? Jheald (talk) 13:56, 20 February 2017 (UTC)

"reading across" between different hierarchies

The VoB data keeps different hierarchies very separate, cf

• User:Jheald/todo/VoB/hierarchies

But there may well be useful queries that we want to do that "read across" the hierarchies, e.g.

• Which ecclesiastical parishes were in a particular civil sub-registration district in a particular year of the 19th century ?
• Which of those sub-registration districts covered a particular present-day unitary authority ?
• What proportion of civil parishes contained multiple ecclesiastical parishes ?

What statements/properties should we put in place to facilitate queries like this?

One difficulty is that if we use located in the administrative territorial entity (P131),

• there are users who may remove such relationships as anachronistic

cf the "end date" someone has put on the statement Edinburgh (Q23436)located in the administrative territorial entity (P131)Scotland (Q22) (not after 1975, apparently)

• (and, to be fair, it does make extracting particular hierarchies more difficult).
• it greatly increases the nesting problem above.

But how else to do it? Jheald (talk) 17:43, 20 February 2017 (UTC)

Note: given that there is now a constraint on located in the administrative territorial entity (P131) that the things it connects should be "contemporary", ie both should have coexisted at some point in time, avoiding anachronism, I have proposed a new property "located in present-day administrative territorial entity", to relate historical administrative units to those in the present. That might help with this a little. Jheald (talk) 08:57, 21 February 2017 (UTC)

item overloading

The VoB model is very strict in avoiding item overloading. To start with, it has separate items for all places and all areas -- there is no overlap of function. (Even for counties, which one might think might only exist as areas, there are separate 'place' pages)

Secondly, (with the exception of areas at the parish level) there are different items for areas if they exist in different hierarchies, which are kept rigidly separate.

• It is therefore easy to reconstruct a hierarchy just with a single "isPartOf" relation, without needing any qualification
• The 'place' items act as the main entrypoint for casual readers, with links to all the areas in all the different structures associated with that place.

In contrast, our practice on Wikidata has (so far) been almost to the other extreme: to be extremely lazy about creating new items, mostly doing so only if forced to do so if there are wikis that differentiate between them.

The extent to which different UK administrative items are overloaded/multi-roled can be seen via the 'co-classes' query links for each class of administrative item on these pages for England, Scotland, Wales, and Northern Ireland; and also this page for different kinds of settlements.

Parishes / villages

• Currently of 7269 items for civil parish (Q1115575), over 5500 (ie all except about 1700) are also items for villages (5064), towns (369+), hamlets, etc.

It might well make sense to have separate items for the parish and its primary settlement; but then the question arises, which item gets which statements (and which sitelinks). It might make sense to connect wiki articles to the settlement, infoboxes could then easily enough pull from the parish as well. But Commons categories as they are at the moment belong on the parish, because they typically include images (often imported from Geograph) from outside the main settlement. However, this risks breaking the P373 link between the two, which would directly affect links shown on the wiki -- they would appear to have no Commons categories anymore. Does one accept this, and just have templated "patch"-style links on the wikis, out of sync with the sidebars? Does one allow multiple P373s to point to the same Commons cat? Not sure.

A split will also involve a fair amount of work updating P131s, of which there are a great many -- almost all listed buildings have one, that were bulk uploaded for Wiki Loves Monuments.

Cities, boroughs and districts

• Going up to larger-scale settlements, a useful set to look at is present-day cities tinyurl.com/jptr2vz. "City" was never an administrative status, so this contains entities with a range of different current administrative arrangements. Nevertheless, a vaguely typical progression (in England) might have been for an ancient borough to have been re-established as a "municipal borough" in 1835, a status renamed "county borough" in 1888, then to have been re-established as a "district" in 1975 (sometimes incorporating a rural area around the city), which in the latest 2009 reorganisation may or may not have continued as a distinct administrative unit; alternatively administration may now be in the hands of a wider unitary authority.

Typically a civil parish was also created to correspond to the territories of such a municipal borough / county borough in the wake of the en:Local Government Act 1894 by amalgamation of older parishes (and sometimes later enlarged with them). These persisted until 1974 when these boroughs were replaced with new districts, and their old civil parishes were abolished, leaving the areas "unparished".

The treatment on Commons can be interesting, because (particularly to categorise images from en:Geograph), Commons has a very systematic division of the UK into entities at different scales -- typically distinct counties, districts and then parishes. For Commons it was important to place images geographically, so each entity has a very clear area associated with it. (So even when Commons categories for settlements are not current administrative areas, the settlement category does not represent a "pure place", but has its own specific area identified with it).

• Once class of such cities, then, is where there is a territorial distinction between a city and a current administrative district named after the city. The following cities in England for example fall into this class:

Bradford, Canterbury, Carlisle, Chelmsford, Chichester, Coventry, Gloucester, Lancaster, Leeds, Lichfield, Newcastle, Peterborough, Preston, Salford, Sheffield, St Albans, Sunderland, Wakefield, Winchester, Worcester, and York.

Both Commons and most larger wikis have separate pages for the two, with Commons typically identifying the "city itself" with the now-unparished area corresponding to the former municipal borough.

• A second class is where there is now no district-level administrative area specifically identified with the city. This is now the case in England for the following, though all were formerly municipal boroughs;

Bath, Chester and Durham

and also for these cathedral cities (although these remained civil parishes, so do have parish councils, which may style themselves as city councils).

Ely, Hereford, Ripon, Salisbury, Truro, and Wells

• As a final third class are the cities where the city's current administrative district is co-terminous with the city proper. Cites in this class include:

Bristol, Cambridge, Derby, Exeter, Kingston, Leicester, Lincoln, Liverpool, Manchester, Norwich, Nottingham, Plymouth, Portsmouth, Southampton, Stoke-on-Trent and Wolverhampton

Commons does not distinguish between the settlement and the administrative unit for these. We, on the other hand, do have separate items, because there are separate pages on ceb, sv, and (sometimes) pl wikis; though not en-wiki. One consequence is the "district" items for these places in the P131 sequence have no Commons category to correspond to.

Starting instead from the side of administrative areas, the metropolitan boroughs tinyurl.com/z7kzye9 all have distinct Wikidata items; most (but not all) have corresponding Commons categories and en-wiki articles (the exceptions being ones that are co-terminous with cities or towns). The same is also true for non-metropolitan districts: tinyurl.com/hz48vub

As regards historical units, the relevant categories on en-wiki already include a fair number of distinct articles, with corresponding distinct items here, eg en:Category:Municipal_boroughs_of_England, en:Category:County boroughs of England, en:Category:Urban districts of England, en:Category:Rural districts of England, en:Category:Former non-metropolitan districts, en:Category:Former non-metropolitan counties, so there seems no objection to making items in these classes. VoB combines entries for urban districts, municipal boroughs and county boroughs, when they reflect essentially the same entity, albeit with a legal status that evolved over time. The post-1974 districts and boroughs outside London all (I think) involved significant amalgamations of multiple earlier entities, so it makes sense for them to be different items.

Counties

• Typically, for many "shire" counties in England we currently may only have a single item, which may carry the roles of historic county of England (Q1138494) and ceremonial county of England (Q180673) and non-metropolitan county (Q769603) -- indicating that it existed pre-1974; it is a county for present-day ceremonial purposes; and it is a county for present-day administrative purposes.

Some counties (eg Cumbria (Q23066)) that were created in 1974 are not historic county of England (Q1138494)s; others (eg Cumberland (Q23360)) that were abolished in 1974 are only historic counties.

In some counties, since the 1990s/2000s, a part of the county has been separated to be administered separately as a unitary authority. In these counties, a new item has been created for the remainder, and this has been identified as the non-metropolitan county (Q769603). (However the historical progression may have been more complicated than this, as this table on en-wiki notes, as for example in Bedfordshire, where the present set-up was achieved in two steps, with one part of the county being separated in the 1990s, and another in 2009).

Some of the counties created in 1974 have since ceased to exist, eg Avon (Q929902), Cleveland (Q24550), Hereford and Worcester (Q1609779), Humberside (Q1636805). These are marked as non-metropolitan county (Q769603) with an end date, and former administrative territorial entity (Q19953632).

• The "Vision of Britain" website has separate items for Ancient County (pre 1889), Administrative County (1889-1974), Modern (post 1974) County, and District/Unitary Authority.

Does it make sense to split our own items in a similar way?

In many cases (IMO) probably not. Consider eg en:Norfolk. Yes, over the years, its precise legal status and governance may have changed; but ut essentially has historical continuity as much the same entity since before the Norman Conquest -- as we reflect by treating its entire history in a single article on Wikipedias, as a single continuous entity; with a corresponding category on Commons. It makes sense to have a single item that reflects this.

One could make a case that it might make sense to have an additional item for the (1889-1974) administrative county, which for administrative purposes excluded the city of Norwich, which over that period had self-governing "county borough" status. (While remaining part of the whole of Norfolk for ceremonial puroposes). But this seems to be a finer distinction than is made by VoB, which simply has Norwich as part of the "administrative county of Norfolk" for its whole history, from 1835 (its initial establishment as a "municipal borough") through to 1974. (One can also note that the boundary map of the "administrative county" includes all of Norwich). It seems that the VoB split may be more to define some clean non-overlapping hierarchies, which may more reflect how particular historical statistics that VoB is interested in could be dis-aggregated, rather than necessarily being absolutely faithful to history. Our priorities may be different, with more value on grouping things together, rather than splits which can be hard to navigate, and hard to map onto the existing structures that have evolved on the wikis and on Commons.

(Another example of a questionable VoB split can be seen in its treatment of modern London boroughs, where there are eg separate area items for the Royal Borough of Kensington and Chelsea for 1965 to 1974 and 1974 onwards. Again, this division may make sense for VoB, because of the different statistics it may have available for the two periods (compare the rows in the table for Kensington on the "units covering this place" tab); but it's not clear that it makes sense for us.)

Yes, objections to retaining a single principal item for the whole history of some traditional counties could be made on the basis of changes that did happen to them in 1974 -- see for example the before-and-after map of eg en:Lancashire for one particular case. But on the other hand, many counties have seen a history of changes, not just in 1974, for example en:Evolution of Worcestershire county boundaries since 1844 -- so changing boundaries are something we have to find a way to deal with anyway. Jheald (talk) 00:06, 1 March 2017 (UTC)

more on current patterns of item overloading and nesting

• Pages for England, Scotland, Wales, and Northern Ireland give a detailed look at how items in different UK administrative levels (current and historical) have been treated for example with regard to overloading, wiki articles, and Commons categories. In particular, the 'list/analysis' column of queries lists the items in each class, together with any other classes each item is a member of, wiki articles, Commons categories, and administrative units it is in the territory of.
• Here also is a query written to try to reproduce the current English structure of regions, counties, unitary authorities and districts for comparison with en:Subdivisions of England, followed by comments discussing the structures and relationships assumed:

Some comments on the query

• The query tries to generate the hierarchy from region of England (Q48091) -> ceremonial county of England (Q180673) -> Current county-level administrative territory -> Districts, to be similar to that at en:Subdivisions of England.
• At each stage, the strategy is to start at the current level, see what items are located in the administrative territorial entity (P131) that level, then restrict to those of an acceptable type (since all sorts of other items might be P131 the same item).
• Overloading of items is present: for example Cumbria (Q23066), Surrey (Q23276), West Sussex (Q23287) are considered both county-level administrative territories and ceremonial counties - the boundaries exactly coincide. Greater London (Q23306) is an administrative entity, and a ceremonial one, and a region. In other cases there are different items, even where counties have the same name, if not all of the ceremonial county falls within the area administered by the county county.
• Multiple values of instance of (P31) have been used to indicate multiple statuses. But this comes with a risk, that users will may decide that one of the roles is to be considered the preferred primary role, ie by giving one value of instance of (P31) preferred rank -- so that this value, rather than any of the others is chosen to appear in the infobox. (Infobox templates on some wikis do allow classes of values to be preferred, or to be blacklisted, for the purpose of what to show in the template. Nevertheless when a scheme requires a property like P31 to have multiple simultaneous values, the risk should be considered that one may be made 'preferred'. Similarly for P131.
• As a result, the form p:Pxxx/ps:Pxxx has been consistently used, instead of wdt:Pxxx. But this can be less efficient (the optimiser finds it harder to optimise); and may be more difficult to hand-optimise (since, with a directive to turn the optimiser off, Blazegraph may think it is being told to approach all of these compound relations by joining from the left first [to be confirmed]).
• The query doesn't (at the moment) include many explicit checks on whether entities still exist; or whether any of the relationships have an end time (P582) qualifier. For ceremonial counties, it does check to exclude any that are former administrative territorial entity (Q19953632). For districts it relies on all that existed after 1974 being instances of non-metropolitan district (Q1187580), metropolitan borough (Q1002812) or borough of London Region (Q211690); and that vice-versa all such districts continue to exist, except for where there is now a unitary authority area in England (Q1136601) -- so if a unitary authority is returned at the county level of the search, then any districts are locked out.

It would be an interesting challenge to develop the query to work for arbitrary earlier dates, probably requiring work on both the query and the data.

• Even without further historical elements, the structure shows deviations from pure nesting. For example, North East Lincolnshire (Q1342914) is part of Lincolnshire (Q23090), part of which is in Yorkshire and the Humber (Q48063) and part in East Midlands (Q47994). But none of North East Lincolnshire (Q1342914) is in East Midlands (Q47994). So at the moment this is locked out by hand. But perhaps a relation is needed, to express that X is not (not any of it) located in the administrative territorial entity Y.

Alternatively, the code could perhaps be tweaked so that if an item has one located in the administrative territorial entity (P131) <item> with qualifier object has role <higher level administrative type>, then it could know to lock out the rest. Jheald (talk) 19:39, 25 February 2017 (UTC)

Ancient and modern cities

I am interested in modeling the relationships between ancient cites and their modern successors where we have separate items due to Wikipedia articles.

Gades (Q3094251): Phoenician city, precursor to Cadíz, Spain > Cádiz (Q15682): municipality and capital city of the Province of Cádiz, Spain

Aquae Sulis (Q622992): town in Roman Britain on the site of Bath, England > Bath (Q22889): city in Somerset, England, United Kingdom

Would located in the present-day administrative territorial entity (P3842) be a reasonable property to use here? Some are currently tagged with <followed by/follows>, although strictly those properties should only be used as qualifiers (a constraint that is currently ignored in ~83% of occurrences). - PKM (talk) 01:42, 15 April 2018 (UTC)

I will add this to the proposed best practises on the project page. – Susanna Ånäs (Susannaanas) (talk) 16:23, 5 April 2019 (UTC)

This is an interesting question and I wonder if there's any relationship to whether they were continuously inhabited or not? For example, is the Roman town truly a different 'thing' from the English town, or just how the same settlement over time has been referenced? Is ancient Rome the same thing as modern Rome? Jeffme (talk) 18:45, 26 December 2020 (UTC)

What are the best modelled items for your areas of interest?

Hi all

Over the past few months myself and others have been thinking about the best way to help people model subjects consistently on Wikidata and provide new contributors with a simple way to understand how to model content on different subjects. Our first solution is to provide some best practice examples of items for different subjects which we are calling Model items. E.g the item for William Shakespeare (Q692) is a good example to follow for creating items about playwright (Q214917). These model items are linked to from the item for the subject to make them easier to find and we have tried to make simple to understand instructions.

We would like subject matter experts to contribute their best examples of well modelled items. We are asking all the Wikiprojects to share with us the kinds of subjects you most commonly add information about and the best examples you have of this kind of item. We would like to have at least 5 model items for each subject to show the diversity of the subject e.g just having William Shakespeare (Q692) as a model item for playwright (Q214917), while helpful may not provide a good example for people trying to model modern poets from Asia.

You can add model items yourself by using the instructions at Wikidata:Model items. It may be helpful to have a discussion here to collate information first.

Thanks

John Cummings (talk) 15:13, 17 December 2018 (UTC)

I hope we will be able to provide some insights after a bit of discussions. I will work in this space in the coming weeks. – Susanna Ånäs (Susannaanas) (talk) 16:26, 5 April 2019 (UTC)

@Susannaanas: Did anything ever come from this? --Oravrattas (talk) 14:25, 7 June 2020 (UTC)

Help with a Historical Place

Hi! I'm newish here and I'm working on a place from prehistory. Could someone come by and look at my work on Q223385 (Cueva de las Manos), and provide some help/feedback? Thank you so much! Best, Tyrone Madera (talk) 23:59, 7 July 2021 (UTC)

significant events for buildings

I am interested in timelines for fortifications, and how to use inception (P571), significant event (P793), service entry (P729) and service retirement (P730) . What significant event (P793) events should I use for drawing up plans, construction start, construction end, remodelling, stopping use as a defensive structure. Generally fortifications don't disappear, but are either repurposed as museums, business parks, dwellings or museums, what standards do you have for that? What if a castle had a 1080s motte and bailey, but all we see now is a 13th century structure with 19th century additions.

My inclination would be to have inception (P571) to duplicate construction start. service entry (P729) and service retirement (P730) are generally used for equipment, but could be used for buildings defining when they were used for their original purpose, but should they wait until the building is fully operational (an unfinished fort is still defensible), and how do you define an end (for a fort, its military purpose ends when the army move out, but like lots of buildings they get repurposed, and a castle's military use just fades away). Perhaps inception (P571) is the only useful date for a building. Vicarage (talk) 09:41, 29 November 2022 (UTC)

Property proposal/British Listed Buildings Online ID

I'm looking for opinions on Wikidata:Property_proposal/British_Listed_Buildings_Online_ID Vicarage (talk) 08:07, 7 February 2024 (UTC)