Elliptical Toponyms

General talk
, , ,
1

Hi everyone!

I recently wrote a short note about elliptical toponyms—place names where part of the name, often a generic term, has disappeared.

Here are some examples of such names in Bulgaria, where the community tends to drop the generic part of the name. More historical and contemporary examples can be found in my diary post.

Full Name Entered in name tag Context
ulitsa “Zapadna” Zapadna “Zapadna” means “Western.” A common directional street name.
ulitsa “Vladaiska” Vladaiska Named after the Vladaiska River, which flows through the heart of Sofia.
ulitsa “Uiliam Gladston” Uiliam Gladston Named after William Gladstone, the British statesman

I focused specifically on names that, in their current form, have not diverged so far from their original shape that this element would be perceived as redundant, yet in many sources they are presented in a truncated form. This can happen because people don’t notice it, or because in many databases the generic term and the main name are separated into two fields, which can break both the syntax and meaning of the name.

I’d love to hear from interested mappers if you have examples of elliptical toponyms, and what the situation is like in your area. For older or official names, the shortened form might be normal, but for local or unofficial names, it can be more flexible and nuanced.

I find this topic interesting because it connects linguistics, mapping, and computer science. I want to explore it more and maybe describe it in detail on the wiki. I might have missed some existing materials, so if you know any articles or discussions, I’d really appreciate your guidance.

Another reason I care about this is to help mappers understand these names better—their meaning, structure, and usage—so that when possible, names are mapped in their original form, not just how they appear in gazetteers.

There is also a related discussion about missing descriptive terms in object names, which intersects with this topic.

2 Likes
역 이름 등에서 '-역' 같은 유형항을 빼지 마십시오
Can we tag an infinite number of alternative names for a place?
2

Something I’ve been curious about is that some Slavic languages typographically distinguish between the generic and specific by putting the specific in quotation marks, not only in road names but also in the names of points of interest and buildings. Are there grammatical reasons for this? Could it influence the choice to omit the generic from name=*? As we discussed in this footnoted thread, East Asian languages tend not to separate the generic and specific in writing or signage. We need to strike a balance between reflecting real-world practices and ensuring that the data is usable in a global context, because some use cases like search cannot inherently be limited to a particular country or culture.

Similar to the Slavic practice, English used to capitalize the specific but not the generic. Over time, this distinction has disappeared for some feature types like roads, while other feature types like place names have instead tended to drop the generic. Exceptions abound, often creating problems in addressing systems. Many U.S. cities have a street named “Broadway” without any generic. In Spanish-influenced California, El Camino Real is such a famous road that you can say “El Camino”, dropping the specific, and any Californian knows what you’re referring to. Some less famous streets have been named after it, as just “El Camino”. The Spanish left lots of toponyms ending in -ero. English speakers appended a generic to some, like Divisadero Street, but appended a definite article to others, like The Embarcadero – a generic specific.

Sticking with California, the official statewide system for naming bridges uses several generics, such as “Overhead” if a bridge over a railway is named for the street it carries, “Undercrossing” if the bridge is named for the road passing under it, and “Separation” if it’s named for two roads that would otherwise intersect. But if a bridge is named for the waterway it crosses, then the name lacks a generic. Thus “Guadalupe River” can refer to a number of road bridges. “River” is only a generic in the name of the river itself.

1 Like
Reviving the area:highway Rendering Discussion: A Strategic Case for QA and Data Integrity
3

I don’t think it has to do with “Slavic” generally – there aren’t significant grammatical differences between Slavic languages that would justify that. What is at play are likely different orthographic traditions, and mutual influences that resulted from more or less shared cultures.

I would ascribe (over)use of quotation marks to Russian first and foremost, and on the basis of shared cultural traditions, it would extend to Belarussian and Ukrainian, and then to Bulgarian and, to a lesser extent, Serbian. Speaking of my native Serbian, the latest edition of the official orthography manual specifies that, paraphrasing, “quotation marks around proper names should be used sparingly, mostly in cases if a confusion may arise whether the reference is to a generic or a proper name”. So, it’s not very different from English. Yeah, I do find taggings such as name=Restoran "Dunav" which I relentlessly trim, and that’s largely the community consensus. Can’t tell for sure about West Slavic languages, but I haven’t seen much use of quotation marks there, so it’s likely similar as in Romance and Germanic languages.

3 Likes
4

In the US, it’s highly variable. In Spokane, it’s common to drop both the directional and the generic from street names, both in speech and on signs. For example, “North Division Street” is often reduced to “Division” because “South Division Street” is simply the southern extension of that same street on the address grid, and like nearly every other street in the metro area, the name is unique.

The Seattle area is a very different situation. When I lived there as a child, it was drilled into me to always include the directional and generic when giving my address: the use of a county-wide numbering system means that North 124th Street runs parallel to and about ten miles north of South 124th Street, and they both intersect with (extensions of) 124th Avenue East.

2 Likes
5

At least in Poland, the standard is to also drop the generic (ulica) on OSM but here in Germany, we always write down the full name in part because of our composition rules (many streets are called ”-straße“ e.g. Eichstraße, Wallstraße) but even where ”Straße“ is its own word (due to adjectives e.g. Breite Straße, Frankfurter Landstraße), we still keep ”Straße“ around for consistency.

2 Likes
6

Thank you for the examples. As I see it, Broadway, El Camino, and Division are a few isolated cases of names that have undergone ellipsis due to their relatively high prominence and cultural significance. Ukraine has a comparable example in Khreshchatyk, the central street in Kyiv, which everyone knows and typically refers to without any generic term.

I may be mistaken, but it seems to me the situation could actually be the reverse: quotation marks might emphasize a certain convention, a particular way of conceptualizing such names in the language—as fully self-contained units, alongside which the generic term may even feel redundant.

This assumption is not unfounded. For instance, only recently did I start looking at settlement names in Ukraine through the lens of “generic-term ellipsis,” and I noticed that a large share of Ukrainian village and settlement names are in fact truncated in this way. Syntactically and semantically, they still presuppose a generic term, since they originate as adjectives in a structure like “specific (adjectival modifier)” + “generic (head noun).” For example, adjectival settlement names like Vilne [selo], Stare [selo], or possessive forms like Shevchenkove [selo] and Volodymyrske [selo] sound quite unnatural to me, as a native speaker, when combined with the explicit generic term in that order. By contrast, street names formed in a similar way—Vilna vulytsia, Stara vulytsia, Shevchenkova vulytsia, Volodymyrska vulytsia—are entirely normal in their full forms.

Feels weird (with generic term) Completely normal (with generic term)
Vilne [selo] (Free [village]) Vilna vulytsia (Free Street)
Stare [selo] (Old [village]) Stara vulytsia (Old Street)
Shevchenkove [selo] (Shevchenko’s [village]) Shevchenkova vulytsia (Shevchenko’s Street)
Volodymyrske [selo] (Volodymyr [village]) Volodymyrska vulytsia (Volodymyr Street)

The age difference between these toponyms is relatively short—roughly a century. So it seems that roughly a hundred years may be enough for such ellipses to become fully normalized.

In other words, the fact that truncated toponyms are widespread in countries like Poland, Bulgaria, and others (it would be interesting to identify which ones exactly) may reflect how these names are actually perceived and used. In Ukraine, this phenomenon applies, broadly speaking, mainly to rivers and settlements, whereas in other countries it extends to a wider range of name types.

What concerns me are the borderline cases, where the situation “oscillates” within a particular country or linguistic environment—that is, where both the full and the truncated forms are equally acceptable. In my view, this is precisely where local OSM communities, linguists, and those designing official registries and databases should step in and gently steer practice toward preserving and recording toponyms in their full forms. I believe this is something we are capable of achieving.

@Duja, I would like to ask you, as a native speaker, how elliptical street names are perceived in Serbia—do the full street names feel “unnatural,” and do my reflections make sense?

1 Like
7

The thread you linked in the footnote eventually digressed to the topic of railway station names. English and apparently in other Western European languages use the generic “station” in some contexts but not others. Map labels, running in boards, and timetables tend to omit it, but it does appear in prose, where “station” is less likely to be repeated a bunch of times in the same breath.

This has caused so many edit wars and difficult discussions on the English Wikipedia as to railway station article titles: “Springfield”, “Springfield Station”, “Springfield station”, “Springfield railway station”? The global English Wikipedia community eventually “agreed to disagree”, delegating the decision to each national community of rail editors, who then proceeded to bicker some more. When MapTiler proposed to feature the OpenMapTiles layer on the OSM homepage, one of the sticking points was that the Wikidata fallback cluttered the map with generics on railway station labels, based on these Wikipedia naming conventions.

For street names in English, the full and truncated forms can both be useful, even in places like Seattle where the generics and directionals are much more significant. Knowing the structure of a name would enable data consumers to avoid overabbreviating the name and allow them to format the specific more prominently than the generic, as seen on standard U.S. street name signs and while navigating in Google Maps.

Some local communities do use name:prefix=* and name:suffix=* for specific purposes, but these subkeys only allow us to isolate the specific by process of elimination. A directional can be either a prefix or a suffix, and there can be multiple suffixes. A name:specific=* would be more elegant for this purpose, especially in cases that lack a generic.

1 Like
8

And for France: we always use the full name in street names (as we have several prefixes for that, otherwise you couldn’t make the difference between rue (street) and impasse (dead end) for instance, both being in the same place).
For POI, we tend to avoid things already well described by tags, so no “Gare de” (station), no “Restaurant” (I guess you understood French here).

For rivers, we don’t say rivière, fleuve but we do say ruisseau (so small streams get the generic part or not, depending if it makes sense without), but we always use an article: “La Seine” for instance.
We have some exception, for instance we can hear “La rivière de Landerneau” (Landerneau being the town where tidal influence starts/stops), but on maps you’ll see L’Élorn".

How natural is something like Mer Noire without the Mer for you?

It simply makes no sense, if you are speaking from “La Noire”, people will think you’re speaking about a black cow named “La Noire” (well “La Noiraude” would be more common). Never about the Black Sea.

2 Likes
9

I completely agree regarding streets and POIs. I want to support my assumption that the structure of a toponym should be decisive in determining whether its name can be given in a shortened form or not. How common are toponyms in France where the specific part is an adjective agreeing with the generic noun? In my country, almost half of them are like this… But I see that in OSM, in many Eastern European countries, they are often “trimmed.” :thinking: How natural is something like Mer Noire without the Mer for you?

10

Can’t speak for France, but in Vietnam, there’s a national orthographic standard that one must capitalize e.g. “Sông Hồng” (literally, Red River) and “Biển Đông” (East Sea) but not necessarily “sông Mê Kông” (Mekong River) or “vịnh Hạ Long” (Halong Bay), depending if the specific can be used on its own attributively in phrases such as thuyền Hạ Long (boat on Halong Bay). Ironically, I recall that French maps are more likely than Vietnamese maps to label the large southern city as just “Hô Chi Minh”, omitting “Ville”. It must be spelled out and capitalized as “Thành phố Hồ Chí Minh” in the local language.

1 Like
11

I tried looking up places in Vietnam. It seems that most rivers and lakes are listed with a generic term. I want to ask about these village names: Mới, Năng, Thái, Trang, Thịnh, Vàng, Xoan. Are they adjectives, and could they be elliptical names that work fine without a generic term?

12

We’re putting generics in natural feature names mainly for internal consistency and consistency with other languages. After all, if we follow the real-world standard and only include “sông” where essential, then will a data consumer know to prepend “sông” onto only the names of rivers that don’t strictly need it? Either way, it’s unfortunate that data consumers can’t easily identify the specific without language-specific heuristics.

Parts of speech in Vietnamese can be fluid. Some of these names are adjectives (mới = new), some are nouns (xoan = bead tree), some can be either nouns or adjectives (vàng = gold or yellow), and some may be names borrowed from indigenous languages respelled using the Vietnamese alphabet. The part of speech matters less than whether the word is unique enough to be understood as a place name.

Traditionally, a Vietnamese place has two names: a Sino-Vietnamese “character name” used in formal contexts (tên chữ) and a demotic name used informally on the street (tên nôm). A typical character name is disyllabic, so it’s likely to be understood without a generic, but we still use a generic when inserting it into a sentence. Urban neighborhoods and rural villages are more likely to be known by their demotic names, which often take the form of a generic plus a monosyllabic specific. Consider the whole thing to be a specific.

In Sóc Vàng, sóc is the word for an indigenous Khmer hamlet, and vàng means gold or yellow in Vietnamese. If you say Vàng by itself, no one will understand it as a place name. Whoever mapped this hamlet as name=Vàng place:VI=sóc appears to have overanalyzed its name, because the complete descriptor would be “ấp Sóc Vàng”, ấp meaning “hamlet”. Maybe they flubbed something in a spreadsheet before importing. I’m sure they wouldn’t have made the same mistake with Sóc Trăng, a well-known city and former provincial capital. Its name literally means “moon hamlet”, but it’s actually a corruption of a Khmer name meaning “silver storage”. This is of course the stuff of etymology, not the proper method of naming a feature in OSM.

Mới was imported many years ago from GNS, which has very problematic coverage of Southeast Asian place names. Over the years, we’ve grappled with mojibake, bogus entries, and very questionable names. The same user also imported hamlets from more reputable Vietnamese government datasets that have since vanished. Some of us have tried many times in vain to track down the original sources or alternative sources so that we can figure out the true names and locations. This is a difficult task in a country which is only officially divided into provinces and communes. Anything more local than that is much less formal. (Imagine if a country half the size of Ukraine were only divided into oblasti and raiony but nothing beyond that. Imagine that officially there’s no such thing as a village or neighborhood, yet those are everywhere.)

2 Likes
Can we tag an infinite number of alternative names for a place?
13

Thank you for the excellent and detailed explanation. We have also relatively recently begun a reform of the administrative system, which, however, was never fully completed due to more pressing challenges that arose for the government. The GNS data for Ukraine are, to put it mildly, not of the highest quality either, so I can imagine the confusion that importing them might cause.

I can draw some parallels between old and current generic terms in Vietnam and Ukraine. For example, the term khutir, translated as “hamlet,” is no longer in official use. Today, there are officially misto (a large town), selyshche (a small town), and selo (village). These three terms are explicitly distinguished in our system under name:prefix. Some settlement names are syntactically complete and still include the old archaic category (such as sloboda [free settlement] or khutir [hamlet]): for example, Kozatska Sloboda, Rokytna Sloboda, Veselyi Khutir, Zhurbyn Khutir.

However, many names have naturally lost their generic element over time and are now used in an elliptical form. This is quite apparent to me as a native speaker from their adjectival or possessive suffixes. You don’t even have to look far—for example, in our capital, Kyiv, the suffix -iv is possessive, meaning that Kyiv was once grammatically aligned with a common noun… but that was about 1,000 years ago :slight_smile:

2 Likes
14

Like in Ukrainian (I presume), ulica is the primary term for a street, and it is normally omitted in speech and even writing, whenever the context is clear. While it is normally written (usually in a tad smaller font) on street signs, Serbian community decided to systematically omit it from the map. An exception was made only for officially numbered streets, such as 22. ulica, but those are generally rare. Spelling or pronouncing ulica is normally done only to introduce it in context (e.g. in news, such as U Temerinskoj ulici se desila saobraćajna nezgoda…) but would generally feel stilted or formal otherwise.

There are few other terms in use (trg ‘square’, bulevar ‘boulevard’, put ‘road’, avenija ‘avenue’) and those are normally written and pronounced. For example, in my home city, Temerinska [ulica] extends into Temerinski put, and similarly Futoška [ulica] into Futoški put.

For some reason, the neighboring Croatian community, despite the same language structure, decided to keep the word ulica explicitly written in OSM. Personally, I find that unwieldy, but they may have some reasons I’m not aware of. I’ve noticed that they tend to have more streets that are “toponymic” in nature (a casual look at e.g. Ilok shows Staklenac, Ribnjak or Turska skela, that may not have ulica attached at all).

Such elision does not normally happen with any other term. Lukino Selo and Staro Selo are only ever called with full name. Some settlement names which are structurally possessives have never had an attributive attached, so the only existing form is short: Čelarevo, Ratkovo, Kljajićevo.

2 Likes
15

Here in Greece, the situation with road names is pretty much the same as @Duja described above. For street, we omit the word from the map because that how it works in everyday language. Besides the street signs, you would see the word “street” (=οδός) mostly in news and official documents. Even for addresses, businesses don’t mention it. But for avenue/boulevard (=λεωφόρος), it’s mentioned even in everyday language, not always but it’s pretty usual. We map those cases with name:prefix

For administrative entities though, unlike most countries, we always use the relevant titles of that entity. For example, we say “Municipality of Ioannina” or “Ioannina Municipality”, and not just “Ioannina” when we want to refer to the municipality’s area. I’ve been contributing at Wikidata a lot the last few years and I see for Greek places there’s a lot of conflating the administrative entities with the settlements by non-Greeks. In Greece, the settlement (settlement/village/town/city) isn’t considered an administrative entity. The lowest administrative entity is called “commune” (=κοινότητα) which may contain more than one settlement., and as mentioned earlier, we always use the title even for that administrative entity.
I’m saying that because I tend to see on OSM that the translation of various admin levels of Greece, don’t contain the title, but only the name. And I heard that at some conference someone non-Greek wanted to propose to omit the “Municipality” part of the “Tempi Municipality” (used as example), for unknown reason. It may be the occasion of how admin levels in other countries are typically called, but that’s not how it works in Greece, and this causes a lot of confusion when fixing data in Wikidata and OSM.

2 Likes
16

That might be exactly the point: this situation—where many European communities systematically drop the generic term “street”—serves as a kind of linguistic marker. Serbian, Greek, Polish, and Bulgarian language communities have gone further in simplifying these toponyms, effectively allowing them to undergo ellipsis, whereas Ukrainian, Croatian, and Belarusian communities generally have not.

This is a completely natural process, as can be seen when comparing street names with relatively older settlement names. The examples you mentioned—Čelarevo, Ratkovo, Kljajićevo—are in fact normalized elliptical toponyms. If I were to add the generic term at the end, I suspect it would sound quite strange and incorrect to you. However, in terms of what the original generic term might have been, it was most likely just selo: Čelarevo selo, Ratkovo selo, Kljajićevo selo.

By the way, the word “ellipsis” itself comes from Greek :) Here is the definition from Merriam-Webster:

… the omission of one or more words that are obviously understood but that must be supplied to make a construction grammatically complete

@jimkats, thank you for describing the situation with these kinds of names in Greece.

1 Like
17

As a matter of fact, all of those are formerly German villages renamed after Partisan WW2 heroes, so selo has never been in the name. It’s just a naming custom over here. In contrast, Savino Selo was also renamed after a WW2 war hero, while Lukino Selo was named after an 18th century nobleman who established it. So elision just does not work in these cases.

Some attribute components of place names are frequently elided in speech: [Sremski] Karlovci, [Stara] Pazova, [Bačka] Palanka (but never in Novi Sad). But those attributes were apparently added later, to “disambiguate” common place names (cf. Frankfurt [am Main]).

18

It seems to me that both in Polish and English mountain names very often match definition of “elliptical toponyms” or maybe it was never ever present.

You have “Mount Everest” or “Broad Peak” or “Ismoil Somoni Peak” in English but also many if not majority of ones without this qualifier.

I also want to note reverse case: “Przełęcz Przysłop” where both names of saddle mean “saddle” - “Przełęcz” is in Polish and “Przysłop” is in Vlach language that was used by distinct ethnic group that since assimilated and disappeared.

To make it more confusing when cartographers were charting mountains they copied/moved names into mountains. This was result of local population caring about naming hamlets, rivers, streams, fields, forests, paths, saddles - more than about naming peaks. While military cartographers cared about peak names. With some hilarious cases, like peak named “Siwakowa Dolina”. “dolina” means “valley” so you get peak named valley.

So, is peak named “Przysłop” (=“Saddle”) an elliptical toponym or not?

What about peak named “Siwakowa Dolina”? (=“Siwakow’s Valley”)

Is saddle named “Przysłop” (=“Saddle”) an elliptical toponym or not? What if it is “Saddle” in an extinct language not spoken by anyone anymore?
What if it would be “Saddle” in language now spoken far-away?

19

Also, behold Node: ‪Spišská Nová Ves‬ (‪26037190‬) | OpenStreetMap - town named “Spišská Nová Ves”

it is funny as Ves == village

Also, there are numerous hamlets in USA called “Something Something City” with name chosen as marketing term during colonization times. Many of them never managed to grow into a city.

20

I can assume that for a native Polish speaker, Przysłop is perceived as a dialectal word, although for the Vlachs it originally functioned as a common noun (?). In this case, both Przełęcz Przysłop and simply Przysłop seem, in my view, acceptable from a structural standpoint. That is, there is no agreement or other internal syntactic relationship within this name. I would not consider the simple form Przysłop to be an “elliptical” name.

In Ukraine, there are many names such as Siwakowa Dolina and Spišská Nová Ves, where the generic term has ceased to indicate the category of the object and has become part of the proper name. Ukrainian orthography prescribes writing generic terms in lowercase, but if they have lost their classificatory function and remain part of the name, they are capitalized. In my opinion, Siwakowa Dolina would also not be considered an elliptical name.

I would like to ask whether I have correctly selected examples of, at first glance, full and elliptical mountain names:

Full forms: Spalena Hora, Pańska Hora
Elliptical forms: Wierchowa, Putyska, Pańskie

If I were mapping peaks in Ukraine and encountered such elliptical forms, I would investigate sources and try to reconstruct the omitted generic term, and in many cases it would likely be possible to restore structurally complete names. However, I would never do this for elliptical settlement names in Ukraine, since it is difficult to imagine more official names than those defined by parliamentary resolutions; the name tag must match the official designation, while alt_name or old_name is another matter.

At the same time, the world is large, and there are many countries where numerous object names have no official status at all. A wiki page about elliptical names, with examples and varying levels of detail, would likely be very useful for emerging local communities.