OpenHistoricalMap
The street that has existed for 150 years that never existed
Posted by Hayleox on February 16, 2025 in English.I found something interesting while working on mapping historical streets in my hometown of Cincinnati. Take a look at Alexander St on J.H. Colton & Co.’s 1855 map, the oldest map I’m aware of that includes the street:
Here’s how S. Augustus Mitchell depicted Alexander St on his 1860 map:
And just for good measure, here it is in Mendenhall 1867, Phillips 1868, Titus 1869, Robinson 1884, Fisk c.1893, Mendenhall 1900, and Mendenhall 1927:
Why am I so interested in finding maps with this old street? Well, mainly because it doesn’t exist! In fact, it has never existed!
Don’t believe me? Here’s modern satellite imagery with elevation overlaid:
(source: Ohio Statewide Imagery Program, USGS contours)
Supposedly Alexander St should appear to the southeast of Rice St (the street in the middle of the picture), running parallel to it for most of its length. I don’t know too much about road construction, but I don’t think a steep hill is a great place to do it.
So how did this road get on all those maps? Well, thanks to this 1883 article in the Cincinnati Enquirer about street name origins, we have some hints:
So Alexander McGrew (who died in 1844) owned this hill, and he declared two streets to exist, and I guess that’s all you had to do back then to get a street named after yourself. It would seem they made it to the county recorder’s office, as I’ve found records of the street being used as the boundary for land plots. And it would seem that all the mapmakers followed suit.
I’ve seen paper streets before, but generally I expect them to be either copyright traps or planned streets in rural/developing areas – not in the urban core of a large city! I’ll take this as yet another lesson that even if something appears on many maps – even ones that promise they’re “from actual surveys” – it might look very different in the real world.
This initially caught my eye when I saw how Sanborn maps marked the street:
I wondered what “Not Opened” means – turns out it means the street doesn’t exist (apparently “unopened right-of-way” is the legal terminology for this).
Despite all the maps these appear on, I think these probably shouldn’t be added to OHM – they wouldn’t be included in OSM, so why should OHM include them? Perhaps they could be added with a highway=proposed tag? Hmmm… if we were going to add them, we’d need to have an end_date… so I’ll need to figure out when they were removed from the map. When might that be?
Well, arguably, never. One last map for you – the official city of Cincinnati map, 2025:1
-
Alright, technically the streets aren’t visible here, just the property boundaries of where the streets used to not-be. Still neat. ↩
Discussion
Comment from Minh Nguyen on February 16, 2025 at 22:55
Without looking, I’m going to venture a guess that Alexander Street is in TIGER too. Many years ago, I had to go through this part of town deleting lots of nonexistent streets that ostensibly climbed up the hill. Maybe this was one of them, or maybe someone got to it before me.
Did you know Hamilton County publishes a substantial dataset of paper streets and vacated streets in the public domain? An import of
highway=proposedandnot:highway=*ways? Peak OHM!Comment from Minh Nguyen on February 16, 2025 at 22:56
Deleting them from OSM, that is. Hard to delete something from OHM that hasn’t been mapped yet. 😛
Comment from BenjaminVis on April 20, 2025 at 08:49
Very nice! In my opinion there is definitely value in retaining as much information as possible from historical sources when ‘digitising’ them. Though this brings into view a discussion about whether one is mapping a best proven version of a previous physical state of development of a landscape or whether one is converting historical sources in a different format to enable alternative (linked) uses. I think any historical GIS ideally would enable both. Or, that is to say, retain traceability of information to the historical source and marking up what of that source is interpreted to be a reflection of a previously existing empirical geographical state or a (historical) construct.
It is not unusual for planned or proposed urban developments to make it onto ‘official’ mapping sources. There is plenty of evidence of historical maps copying information from previously existing maps too (one presumes without putting much effort into checking them). Naturally this also brings into consideration the purposes the maps involved were intended to serve. Explicit navigation doesn’t seem very likely in these instances! The Sanborn maps are (once again) revealed to be valuable and reliable sources, as they clearly recognise the nature of the geographciallly projected situation. Nonetheless, there is an interesting question about the history of the proposed streets and how this is treated administratively and why it never became relevant to update to an empirical reflection of the situation in any of these maps. Indeed, as well, why this proposal still seems to hold relevance presently. What is the legal/land registration status of this?
(Disclaimer: I literally just joined this platform out of curiosity, but am not aware about its explicit purposes or how these are defined. It could well be that the aim is not to digitise all and specific historical sources, but to come to an as reliable as possible interpretation of empirical geographic states of development as we can collective prove to have existed for any point in time. Why, by the way, is the timeline limited to 1825?)
Comment from Minh Nguyen on April 20, 2025 at 15:10
Welcome to OpenHistoricalMap, Benjamin, and thank you for your insightful comments!
You point to a natural tension in historical mapmaking and to some extent in the sort of modern mapmaking that our friends at OpenStreetMap do as well. For me, as a resident of California, I look to the fact that California was once portrayed on European maps as an island – OHM will never show California as an island during that time period, no matter how many contemporary sources say otherwise. As we try to discern the truth, we will necessarily rely on prior art, but we’re fundamentally a work of synthesis, bringing in the best of a variety of sources, including those that say California is not an island and Alexander Street never opened.
For something physical like a landmass or street, at least there’s an objective truth, but we also sometimes grapple with more abstract features. The difference between de jure and de facto boundaries is already fraught enough in OpenStreetMap, let alone OHM, where we have to contend with sometimes outlandish claims made by colonial powers before they even knew what that side of the world looked like. I think the best we can do in many cases is to communicate any caveats about what we show or omit.
Regarding the time slider, 1825 is just a default. We would set the start much earlier, but it would make it more difficult to adjust the date granularly by day. You can adjust the slider’s range using the row of controls immediately above the blue bar. We’ve seen a lot of people overlook these controls; the whole time slider is in need of an overhaul, which could include some improvements to make the range more dynamic.
Comment from BenjaminVis on April 20, 2025 at 15:51
Thanks so much for your elaborate reply and a bit of additional explanation. It seems logical to me, for a platform akin to OSM, that the objective is some kind of synthetic reconstructive representation of historical empirical reality. This does mean that it becomes absolutely crucial for it’s usability that any edit (and potentially multiple versions or readings of a single area or detail) are traceable to source material. Otherwise it would become very difficult to place nay trust in that synthesis later on as a truly critical resource. I have spent very little time exploring, but I dont know how sources are treated and tracked currently. I did notice the difficulty about permission to use material for vectorisation. I’m a little confused as to what this issue extends to, since I’d think that one is free to vectorise whatever is feasible, but not free to the also publish the source material. I’m not an expert on such issues of rights, but it could become hairy pretty quickly. I also don’t think that all archives would properly expect this use of their material.
Regarding de facto and de jure boundaries, I note you’ve opted for very explicitly legal terms. I’ve published on boundaries and used the philosophical concepts of fiat and Bina fide boundaries (of objects or spatial entities, derived from Smith & Varzi) in mapping. For representations of information on maps I continue to find this very useful.
The example you gave of California as an island is interesting. Of course, it becomes very tricky to figure out which actual geographical knowledge could be contained in a document which depicts an entire region as an island we factually know not to be. Nonetheless, it would also be a real pity, historically speaking, if this would render such historical mapping sources essentially unusable. Personally, I feel that it is important to have vectorisations consultable next to geolocated and rectified raster maps for research purposes. In addition there could even be entirely reconstructive vectorisations, which then require marking up as interpretive reconstructions, placing a lot of emphasis on metadata and the process of making. Since I’m carrying out a postdoc project which also touches on all of this (though not a completely conjured up geographical reality, like the California example) I’m still not entirely decided on how best to do this. Part of my interest in OHM was also to find out if it could be the right platform to potentially ‘release/disclose’ (some) of my research data. I’m not clear on that either so far. Possibly because the project wouldn’t respect the parameters set for OHM only an adjusted format of the vector data could become part of OHM alongside other ways of presenting and offering the data.
Good to know I missed something on the time slider. The daring requirement of features, while entirely logical and necessary, is also a tricky one. Very often the precise start and duration of any feature (not even entire spatial objects) is unknown, and historically speaking it would be very useful to have the source that claims the appearance, alteration, or destruction/removal of spatial entities. This becomes even more finely grained when it is the uses that change, not the physical (or legal-spatial) entity.
Comment from Minh Nguyen on April 20, 2025 at 16:46
Yes, unlike OSM, we emphasize citing sources directly on individual features (as opposed to just in changeset metadata) and are gradually building out editor features to facilitate source management.
The situation likely differs from country to country, but copyright can restrict both verbatim republishing and derivation. We’ve been avoiding making full-throated claims of fair use, because of our international audience and the potential for incompatibility with projects that are less scientific or academic in nature. For the most part, we’ve considered out-of-copyright maps and sources other than maps to be safe. (For the latter, we lean on the principle that facts cannot be copyrighted.) Beyond that, we don’t have enough certainty to state a blanket rule yet, but this is all open to discussion.
These are very helpful terms – I may have to borrow them myself. Many of us are essentially self-taught, having gotten into history out of a sense of curiosity or even by accident, so we do need people with expertise to ground our experiments with thoughtful perspective.
There is some wiggle room in terms of mapping competing boundary claims with
claimed_by=*, or mapping representations in order to say they are false, using thenot:*namespace, for example. But we are constrained in the fundamental data model we inherited from OSM: layerless, every kind of data about the world intermingled in one big pot. By even extending this model across time, we somewhat force ourselves to arrive at a definitive answer instead of simply relating every claim from existing sources.We’re thrilled that you’d consider us, even if we don’t end up being the main depository in the original format. Hopefully you’ll be able to find some benefit in a collaboration, whichever form it takes. As optimistic as I am about OHM, the reality is that we’re only one platform in a whole ecosystem of historical GIS. That’s why we strive to enable linked open data through Wikidata, FactGrid, QLever, etc. Platforms such as Pixeum and OldMapsOnline can complement the work we’re doing by presenting original source material in a manner that we’re ill-equipped to.
We have some strategies for modeling date uncertainty and evolution, some of which already have built-in software support. However, there are still some outstanding pain points, particularly when it comes to name changes.
If I haven’t scared you away already, I’d encourage you to start a discussion on the forum or get in touch with the advisory group so we can explore further. Thank you!
Comment from jeffmeyer on April 30, 2025 at 18:09
This is a fascinating discussion! Benjamin - please check your messages, as I’ve sent you a note that’s better shared there than here.
Hayden - thanks also for writing such an interesting and provocative diary entry! : )